ID
int64 1
1.07k
| Comment
stringlengths 8
1.13k
| Code
stringlengths 10
4.28k
| Label
stringclasses 4
values | Source
stringlengths 21
21
| File
stringlengths 4
82
|
|---|---|---|---|---|---|
201 | Transform academic background and job category variables from character to factor | class(expertsample$acadgroup) expertsample$acadgroup<-as.factor(expertsample$acadgroup) class(expertsample$job_cat) expertsample$job_cat <- as.factor(expertsample$job_cat) | Data Variable | https://osf.io/u9hkj/ | expertsurvey_CDRCCS.R |
202 | graphical check for normal distribution of differences support (mean) | diff_support <- bothexpert$BECCS_support-bothexpert$DACCS_support hist(diff_support) | Visualization | https://osf.io/u9hkj/ | expertsurvey_CDRCCS.R |
203 | Wilcoxon tests support (mean) | wilcox.test(bothexpert$BECCS_support, bothexpert$DACCS_support, paired=TRUE) | Statistical Test | https://osf.io/u9hkj/ | expertsurvey_CDRCCS.R |
204 | correlations separately for each condition | data_con <- data[data$Cond == "Control", ] data_ent <- data[data$Cond == "Entitlement", ] chart.Correlation(data_con[ , c("pes","prestige","dominance","benign","malicious","pain")], use = "pairwise.complete.obs", pch = 20, histogram = TRUE) chart.Correlation(data_ent[ , c("pes","prestige","dominance","benign","malicious","pain")], use = "pairwise.complete.obs", pch = 20, histogram = TRUE) rm(data_con,data_ent) | Data Variable | https://osf.io/sb3kw/ | Study2A_analyses.R |
205 | Estimate partial correlation matrices using fiml (partial cors and not glasso b/c different sample sizes leads to different glasso penalties) | net1hi <- cor1.edhi$cor %>% EBICglasso(., n = cor1.edhi$n) net2hi <- cor2.edhi$cor %>% EBICglasso(., n = cor2.edhi$n) net3hi <- cor3.edhi$cor %>% EBICglasso(., n = cor3.edhi$n) net4hi <- cor4.edhi$cor %>% EBICglasso(., n = cor4.edhi$n) net5hi <- cor5.edhi$cor %>% EBICglasso(., n = cor5.edhi$n) net6hi <- cor6.edhi$cor %>% EBICglasso(., n = cor6.edhi$n) net7hi <- cor7.edhi$cor %>% EBICglasso(., n = cor7.edhi$n) net1lo <- cor1.edlo$cor %>% EBICglasso(., n = cor1.edlo$n) net2lo <- cor2.edlo$cor %>% EBICglasso(., n = cor2.edlo$n) net3lo <- cor3.edlo$cor %>% EBICglasso(., n = cor3.edlo$n) net4lo <- cor4.edlo$cor %>% EBICglasso(., n = cor4.edlo$n) net5lo <- cor5.edlo$cor %>% EBICglasso(., n = cor5.edlo$n) net6lo <- cor6.edlo$cor %>% EBICglasso(., n = cor6.edlo$n) net7lo <- cor7.edlo$cor %>% EBICglasso(., n = cor7.edlo$n) | Statistical Modeling | https://osf.io/mj5nh/ | educationnetworksatleast6.R |
206 | 5. Confirmatory Factor analysis To replicate our results without item 1, delete yts_1 from the analyses 5.1 yes/no RESIST | dataCFA1 <- na.omit(data.frame(data$id,data$yts_1, data$yts_2, data$yts_3, data$yts_4, data$yts_5, data$yts_6, data$yts_7, data$yts_8, data$yts_9, data$yts_10, data$yts_11, data$yts_12, data$yts_13)) colnames (dataCFA1) <-c("id","yts_1","yts_2","yts_3","yts_4","yts_5","yts_6", "yts_7","yts_8","yts_9","yts_10","yts_11","yts_12", "yts_13") model1 <- 'EA =~ yts_1 + yts_2 + yts_3 + yts_4 + yts_5 + yts_6 + yts_7 + yts_8 + yts_9 + yts_10 + yts_11 + yts_12 + yts_13' fit1 <- cfa(model1, data=dataCFA1, meanstructure=T, std.lv=T, estimator = "WLSMV", ordered = c("yts_1", "yts_2", "yts_3", "yts_4", "yts_5", "yts_6", "yts_7", "yts_8", "yts_9", "yts_10", "yts_11", "yts_12", "yts_13")) summary(fit1, fit.measures=TRUE, standardized = T) fitMeasures(fit1,c("chisq","df","pvalue","cfi","tli","rmsea", "wrmr",'rmsea.ci.lower','rmsea.ci.upper')) modindices(fit1, sort. = T) | Statistical Modeling | https://osf.io/g2nkw/ | YCAS_Scale_development.R |
207 | get factorscores and include them in origingal data | dataCFA1$factorscore1 <- predict(fit1) data <- left_join(data, dataCFA1, by = "id") data$factorscore1 dataCFA3$factorscore3 <- predict(fit3) data1 <- left_join(data1, dataCFA3, by = "record_id") data1$factorscore3 dataCFA4$factorscore4 <- predict(fit4) data1 <- left_join(data1, dataCFA4, by = "record_id") data1$factorscore4 | Data Variable | https://osf.io/g2nkw/ | YCAS_Scale_development.R |
208 | join behDat and demDat | allData = inner_join(behDat, demDat, by = "userCode") | Data Variable | https://osf.io/wcfj3/ | 1_dataPrep.R |
209 | Standardize Variables zscoring | df_overall = df df_overall$sWM = scale(df_overall$sWM) df_overall$rtWM = scale(df_overall$rtWM) df_overall$rtDivAtt = scale(df_overall$rtDivAtt) df_overall$rtAlert = scale(df_overall$rtAlert) df_overall$sRWT = scale(df_overall$sRWT) | Data Variable | https://osf.io/wcfj3/ | 1_dataPrep.R |
210 | Remove time points above 29, because not everyone attended that often | df_overall = df_overall %>% filter(Time < 29) %>% ungroup() | Data Variable | https://osf.io/wcfj3/ | 1_dataPrep.R |
211 | ensure that userCode is a factor | df_overall$userCode = as.factor(df_overall$userCode) | Data Variable | https://osf.io/wcfj3/ | 1_dataPrep.R |
212 | We can also try to cluster the languages based on the feature values they have. We currently use hierarchical clustering. This will normally also reflect the PCA visualization. The output can also be compared with the family grouping of the languages. generate the clusters | hclust_avg <- hclust(daisy(con_data)) hclust_avg <- hclust(daisy(con_data)) | Visualization | https://osf.io/6hx2n/ | DM_a_k.R |
213 | add squares around the clusters | rect.hclust(hclust_avg , k = 4, border = 1:4) rect.hclust(hclust_avg , k = 4, border = 1:4) | Visualization | https://osf.io/6hx2n/ | DM_a_k.R |
214 | Extracting distances We can extract the pairwise distance between the points of the data set. First, we can visualize the distances in a twodimensional space. Normally, we expect that this visualization matches the output of the PCA. change the content to a distance matrix | distances <- con_data %>% dist(method = "euclidean") distances <- con_data %>% dist(method = "euclidean") | Visualization | https://osf.io/6hx2n/ | DM_a_k.R |
215 | recodes ethnicity (Decline/Other category is set to missing) | sambdif<-samb %>% select(1:19, SAmb_tot, gender, ethnic) %>% mutate(ethnic_fac = rec(ethnic, rec = "1=0;; 6=NA;; else=1")) %>% convert(fct(gender, ethnic_fac)) table(sambdif$gender) # 0=women, 1=men table(sambdif$ethnic_fac) #0=EuroAm, 1=PoC | Data Variable | https://osf.io/ztycp/ | Schizotypal Ambivalence.R |
216 | reorder the list of results to match the alignment of the dendrogram | list.res.reo<-list.res[order.dendrogram(dend1)] layd<-c(10,1) tiff("Figure_hclust2.tif",width=sum(layd),height=11.5,units="cm",res=600,compression="lzw") layout(matrix(1:2,nrow=1),widths=layd) par(mar=c(3.5,1,1.5,14),mgp=c(2,0.8,0)) plot(dend1, cex = 0.8, horiz=T,xlab="Height") | Visualization | https://osf.io/greqt/ | 04_clusters2_tvals.R |
217 | function to read in cleaned data and format it as a tibble that can be used for plotting | read_data_for_plotting <- function(data_path) { dat <- read_csv(data_path, col_types = cols()) | Visualization | https://osf.io/fb5tw/ | plotting_style.R |
218 | create a matrix of relevant item loadings to calculate nonrefined factor (scale) scores | scaleMatrix <- data.frame(data = matrix(EFA$loadings, ncol = 6)) scaleMatrix <- lapply(scaleMatrix, function(x) ifelse(x >= 0.4, 1, ifelse(x <= -0.4, -1, NA))) scaleMatrix <- data.frame(data = scaleMatrix, row.names = colnames(mainData[, c(2:56)])) colnames(scaleMatrix) <- c("Sensory", "CognitiveDemand", "ThreatToSelf", "CrossSettings", "Safety", "States") | Statistical Modeling | https://osf.io/2j47e/ | Cluster analysis non-refined factor scores.R |
219 | get number of patients (from 'table_overview_of_studies.R') | df_long <- left_join(df_long, tab_study_overview %>% select(study_name, N_Patients) %>% rename(Study = study_name) %>% mutate(N_Patients = as.numeric(str_extract(N_Patients, '(\\d*)'))), by = 'Study') | Data Variable | https://osf.io/fykpt/ | 03_table_comorbidities_by_study.R |
220 | add error bars with standard error | plot <- plot + geom_errorbar(aes(ymin=lower,ymax=upper),width=.2) plot <- plot + ggtitle("2a) Average accuracy across four Cue Type conditions") | Visualization | https://osf.io/bfq39/ | Code_LMMs_BestPractice_Example_withOutput.R |
221 | Capture Age for Each particpant | ParticipantAge <- summarize(Summary, count = n(), Age = mean(Age, na.rm = T)) | Data Variable | https://osf.io/2uf8j/ | Negativity R.R |
222 | EVENT VALENCE: Compare Rated Strength of Positive & Negative Events CHECKING FOR OUTLIERS Figure: Boxplot Order Conditions | EventValenceRatings$Valence <- factor(EventValenceRatings$Valence, levels = c("Positive", "Negative")) BoxplotFig <- ggplot(EventValenceRatings, aes(x=Valence, y=Strength)) + theme_bw() BoxplotFig <- BoxplotFig + geom_boxplot(aes(color = Valence, fill = Valence), outlier.size = 2.5, alpha = 0.5) BoxplotFig <- BoxplotFig + labs(x="Event Valence", y="Event Valence Ratings (-50 to 50)") + theme(title= element_text(size=14, face='bold'), axis.title.x= element_text(size=14), axis.text.x= element_text(face="bold", size=14), axis.title.y = element_text(size=14), axis.text.y= element_text(face="bold", size=14), strip.text.x= element_text(size = 14, face='bold'), strip.text.y= element_text(size = 14, face='bold')) | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
223 | Dotplot Valence Figure (Winsorized & ReverseCoded) | EventValenceRatings_MEANFig <- ggplot(data=EventValenceRatings_ByParticipant_MEAN, aes(x=Valence, y=Strength_ReverseCoded)) EventValenceRatings_MEANFig <- EventValenceRatings_MEANFig + stat_summary(fun=mean, geom="bar", position ="dodge", alpha=0.5, aes(fill=Valence)) EventValenceRatings_MEANFig <- EventValenceRatings_MEANFig + theme_bw() EventValenceRatings_MEANFig <- EventValenceRatings_MEANFig + geom_violin(alpha = 0.5, color='grey50') | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
224 | Mixed Effects Model with EventPosition as a Covariate (Linear & Quadratic) Random Effects Structure simplified to converge: ChainPosition removed from ChainID SocialContext removed from Event | Analysis1GLMER <- glmer (Present~ SocialContext.C*Valence.C*ChainPosition.C + poly(EventPosition.C, 2) + (1 + Valence.C | ChainID) + (1 | Event), data=Analysis1, family=binomial) summary(Analysis1GLMER) Analysis2GLMER <- glmer (Present~ SocialContext.C*Valence.C*ChainPosition.C + poly(EventPosition.C, 2) + (1 + Valence.C | ChainID) + (1 | Event), data=Analysis2, family=binomial) summary(Analysis2GLMER) | Statistical Modeling | https://osf.io/2uf8j/ | Negativity R.R |
225 | Mixed Effects Model with EventPosition as a Covariate (Linear & Quadratic) Random Effects Structure simplified to converge: Valence & ChainPosition removed from ChainID SocialContext removed from Event | Analysis3GLMER <- glmer (Present~ SocialContext.C*Valence.C*ChainPosition.C + poly(EventPosition.C, 2) + (1 | ChainID) + (1 | Event), data=Analysis3, family=binomial) summary(Analysis3GLMER) | Statistical Modeling | https://osf.io/2uf8j/ | Negativity R.R |
226 | VISUALISE BIAS (POSITIVE, NEGATIVE) IN EACH SocialContext Substract Mean Positive From Mean Negative Survival Score for each SocialContext Prediction is that Perference for Negativity Decreases first with Communicative Intent and then further with Social Interaction Calculate Means for Positive & Negative Information Calculate Mean Positive Mean Negative for Each Chain Plot Filter | Bias <- filter(NegativityData, Valence %in% c("Positive", "Negative")) | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
227 | Create a Vector of colours based on each Participant's Mean Bias Score (Positive Negative) Use this to Color the dot points by Value | wideBiasMEAN <- mutate(wideBiasMEAN, ColourFig = ifelse(Bias >1, "#006A40FF", ifelse(Bias <0, "#F08892FF", "#95828DFF"))) NegativityBiasFig <- ggplot(data=wideBiasMEAN, aes(x=SocialContext, y=Bias)) NegativityBiasFig <- NegativityBiasFig + stat_summary(fun.y=mean, geom="bar", position ="dodge", alpha=0.5, fill="red") NegativityBiasFig <- NegativityBiasFig + theme_bw() NegativityBiasFig <- NegativityBiasFig + geom_violin(alpha = 0.5, color='grey50') | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
228 | Geom_dotplot with points coloured by Value | NegativityBiasFig <- NegativityBiasFig + geom_dotplot(aes(x=SocialContext, y=Bias, fill=ColourFig), binaxis='y', stackdir='center', dotsize=0.9, alpha = 1, stroke=0.75, colour='grey15') + scale_fill_identity() | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
229 | Add Violin Geom to give indication or data normality | ResolutionCollapsedFig <- ResolutionCollapsedFig + geom_violin(alpha = 0.0, colour='grey50') ResolutionALLPosNegFig <- ResolutionALLPosNegFig + geom_violin(alpha = 0.5, color='grey50') | Visualization | https://osf.io/2uf8j/ | Negativity R.R |
230 | Run IRT model: 3dim model for math, read and scie, no weights Possibility to set seeds and iteration number in the function parameter | mod <- run.irt(pa12_resp_s, items = item.dif) | Statistical Modeling | https://osf.io/8fzns/ | 3_IRT.R |
231 | function to create tick marks on a logscale | log10Tck <- function(side, type){ lim <- switch(side, x = par('usr')[1:2], y = par('usr')[3:4], stop('side argument must be "x" or "y"')) at <- floor(lim[1]) : ceiling(lim[2]) at <- 0:8 return(switch(type, minor = outer(1:9, 10^(min(at):max(at)))[,1:8], major = 10^at, stop('type argument must be "major" or "minor"') )) } if(saveFigures) cairo_ps(file = '../R_Output/Images/OriginalBFScatter.eps', onefile = TRUE, fallback_resolution = 600, width = 7.3, height = 4.16) | Visualization | https://osf.io/x72cy/ | ConfirmatoryAnalyses.R |
232 | Plot: original Bayes factor and replication effect size | plot(original.bf, replication.effectsizes, xlim = c(1, 10^8), ylim = c(-0.2, 1), axes = FALSE, log = 'x', cex = 2, cex.lab = 1.6, pch = 21, bg = c('grey36', 'grey')[replication.outcomes], xlab = 'Bayes Factor Original Study', ylab = 'Replication Effect Size (r)') | Visualization | https://osf.io/x72cy/ | ConfirmatoryAnalyses.R |
233 | take the xaxis indices and add a jitter, proportional to the N in each level | myjitter <- jitter(rep(i, length(thisvalues)), amount=levelProportions[i]/6) points(myjitter, thisvalues, pch=1, col=rgb(0,0,0,.9), cex = 1.5, lwd = 1.5) } graphics.off() rm(myjitter, thislevel, thisvalues, mylevels, levelProportions, nsubjects, subjects) | Data Variable | https://osf.io/x72cy/ | ConfirmatoryAnalyses.R |
234 | Multiple Regression Analysis | multreg.second(Speaking ~ Vocabulary+Grammar+Writing+Reading, corr=correl, n=100) multreg.second(Y~ X1+X2+X3, corr=correl, n=100) multreg.second(Score~ Wordcount+CLI+Commas+Stopwords+Linking+WordsSentence, corr=correl, n=200) multreg(Score~ Wordcount+CLI+Commas+Stopwords+Linking+WordsSentence, data=dat1) lm.out <- lm(Score ~., dat1) | Statistical Modeling | https://osf.io/uxdwh/ | code.R |
235 | Dominance Analysis (using dominanceanalysis package) | lm.cov <- lmWithCov(Score ~ Wordcount+CLI+Commas+Stopwords+Linking+WordsSentence, correl) da <- dominanceAnalysis(lm.cov) print(da) | Statistical Modeling | https://osf.io/uxdwh/ | code.R |
236 | FUnctions for Cleaning Function to generate a table with all observations within a variable and its corresponding count | check_observation <- function (df, column) { require (tidyverse) check <- df %>% group_by (as.vector(unlist(df[, column]))) %>% count () colnames(check)[1] <- column return (check) } recur.collapse <- function(df, n = ncol(df)) { #collapse function combine <- function(x,y){ if(!is.na(x)){ return(x) } else if (!is.na(y)) { return(y) } else { return(NA) } } if(n==1){ return(df[,1]) } else { return(mapply(combine, df[,n], recur.collapse(df, n-1))) } } multi_spread <- function(df, id, key, var){ require(gtools) require(tidyverse) df.list <- lapply(var, function(x){ df.temp <- spread(df[, c(id, key, x)], key, x) # spread 1 variable cname <- mixedsort(unique(as.vector(unlist(df[,key])))) # sort colnames in alphanumeric order df.temp <- df.temp[, c(id, cname)] # rename column names colnames(df.temp)[2:ncol(df.temp)] <- paste(x, colnames(df.temp[,cname]), sep= "_") #add key to column names return(df.temp) }) df.output <- df.list[[1]] # initialise for(i in 1:(length(var)-1)){ # combine matrices together df.output <- full_join(df.output, df.list[[i+1]]) } return(as.data.frame(df.output)) } | Data Variable | https://osf.io/6qej7/ | ECoach Functions.R |
237 | Extract posteriors of effects from Bayesian ANOVA | bestmod <- modBF[1] chains <- posterior(bestmod, iterations=50000, columnFilter="^Subject$") colnames(chains) | Statistical Modeling | https://osf.io/8abj4/ | Exp2.R |
238 | STATISTICAL TEST: Nonparametric Spearman's correlation nonnormalized conditional entropy | cor.test(results$DATE, results$CEDenominationsMotifs, method = "spearman") rdates <- rev(dates) #dates = centuries BCE plot(results$DATE, results$CEDenominationsMotifs, xlim = c(6,4), xaxt='n', xlab = "Century BCE", ylab = "H(D|d)", main = "P2: Conditional entropy of denominantions given designs") axis(1, at = rdates, labels = rdates) | Statistical Test | https://osf.io/uckzx/ | P2_analysis_oldbins.R |
239 | ____________________________________________________________ Plots with H(D|d) nonnormalized conditional entropy Getting statistics: mean, meadian, standard deviation among authorities per period for nonnormalized conditional entropy of denomination given designs | N <- aggregate(CEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = length) MEAN <- aggregate(CEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = mean) MEDIAN <- aggregate(CEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = median) SD <- aggregate(CEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = sd) resultspoleis_summary <- cbind.data.frame(N, MEAN$CEDenominationsMotifs, MEDIAN$CEDenominationsMotifs, SD$CEDenominationsMotifs) colnames(resultspoleis_summary) <- c("DATE","N","MEAN","MEDIAN","SD") resultspoleis_summary$SE <- resultspoleis_summary$SD / sqrt(resultspoleis_summary$N) | Visualization | https://osf.io/uckzx/ | P2_analysis_oldbins.R |
240 | Getting statistics: mean, meadian, standard deviation among authorities per period for normalized conditional entropy of denomination given designs | N <- aggregate(NormCEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = length) MEAN <- aggregate(NormCEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = mean) MEDIAN <- aggregate(NormCEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = median) SD <- aggregate(NormCEDenominationsMotifs ~ DATE, data = resultspoleis, FUN = sd) Nresultspoleis_summary <- cbind.data.frame(N, MEAN$NormCEDenominationsMotifs, MEDIAN$NormCEDenominationsMotifs, SD$NormCEDenominationsMotifs) colnames(Nresultspoleis_summary) <- c("DATE","N","MEAN","MEDIAN","SD") Nresultspoleis_summary$SE <- Nresultspoleis_summary$SD / sqrt(Nresultspoleis_summary$N) | Statistical Modeling | https://osf.io/uckzx/ | P2_analysis_oldbins.R |
241 | aggregate data, zstandardize all predictor variables for the model | recogfreq <- aggregate(cbind(correct, intrusion, new) ~ id+setsize+rsizeList+rsizeNPL, data=recogdat, FUN=sum) recogfreq$zsetsize <- (recogfreq$setsize - mean(recogfreq$setsize))/sd(recogfreq$setsize) recogfreq$zrsizeList <- (recogfreq$rsizeList - mean(recogfreq$rsizeList))/sd(recogfreq$rsizeList) recogfreq$zrsizeNPL <- (recogfreq$rsizeNPL - mean(recogfreq$rsizeNPL))/sd(recogfreq$rsizeNPL) for (sc in 1:length(scales)) { prior <- paste0("cauchy(0, ", as.character(scales[sc]), ")") fixefPrior <- c(set_prior(prior, class="b")) ranefPrior <- set_prior("gamma(1,0.04)", class="sd") | Data Variable | https://osf.io/qy5sd/ | PairsBindingRSS_Stats.R |
242 | Cronbach's alpha (ingroup) | iden.ing.woOutliers.a <- c("ingroup.value", "ingroup.like", "ingroup.connected") iden.ing.woOutliers <- data.all[iden.ing.woOutliers.a] cronbach(iden.ing.woOutliers) | Statistical Test | https://osf.io/9tnmv/ | Exp1a_Election_post.R |
243 | min and max age, mean and sd age, percentage of men and women | minAge = min(df_preQ$age) maxAge = max(df_preQ$age) meanAge = mean(df_preQ$age) sdAge = sd(df_preQ$age) females = length(which(df_preQ$gender == "female")) males = length(which(df_preQ$gender == "male")) other = length(which(df_preQ$gender == "other")) | Data Variable | https://osf.io/xh36s/ | quest_analyses.R |
244 | Check normality with QQ plot and ShapiroWilk test Build the linear model | model <- lm(FW ~ condition, data = df_postQ) | Statistical Test | https://osf.io/xh36s/ | quest_analyses.R |
245 | Question 4: pearson correlation between the different subscales | df_cor <- data.frame(df_postQ$FW, df_postQ$DU, df_postQ$DET) colnames(df_cor) <- c('FW', 'DU', 'DET') res_cor <- rcorr(as.matrix(df_cor)) | Statistical Test | https://osf.io/xh36s/ | quest_analyses.R |
246 | Visualize correlations Insignificant correlations are leaved blank | corrplot(res_cor$r, method = 'number', type="upper", order="hclust", p.mat = res_cor$P, sig.level = 0.05, insig = "blank") | Visualization | https://osf.io/xh36s/ | quest_analyses.R |
247 | Stepwise regression model | step.model <- stepAIC(full.model, direction = "both", trace = FALSE) summary(step.model) step.model <- stepAIC(full.model, direction = "both", trace = FALSE) summary(step.model) step.model <- stepAIC(full.model, direction = "both", trace = FALSE) summary(step.model) | Statistical Modeling | https://osf.io/xh36s/ | quest_analyses.R |
248 | Subset the data set to use base2 instead of base1 for the individuals that have the 2nd baseline Extract values of frogs that have two baselines | base2_frogs <-filter(hormones, condition == "base_02") | Data Variable | https://osf.io/3bpn6/ | os_testosterone_analysis.R |
249 | LMER testosterone by sex and condition with interaction | m1_osT <- lmer(log(t_conc_corr) ~ sex*condition + (1|id), data = hormones) | Statistical Modeling | https://osf.io/3bpn6/ | os_testosterone_analysis.R |
250 | Sex difference in Tlevel Model plot wit sjPlot | plot_model(m1_osT_1, title = "", axis.title = "Fixed effect estimates", axis.labels = c("Time point\n (back home)", "Sex \n(male)"), dot.size = 5, line.size = 2, transform = NULL, sort.est = TRUE, colors = "gs", vline.color = "darkgrey")+ ylim(-0.5, 1) + theme_sjplot2(24) | Visualization | https://osf.io/3bpn6/ | os_testosterone_analysis.R |
251 | now the random slopes part of the model including random intercepts and slopes and their correlation: | pot.terms.with.corr=paste(c(xnames, fe.me[modes[fe.me]!="factor"]), collapse="+")#get all fixed effects terms together... pot.terms.with.corr=paste(paste("(1+", pot.terms.with.corr, "|", re, ")", sep=""), collapse="+")#... and paste random effects, brackets and all that pot.terms.with.corr=paste(c(xnames, fe.me[modes[fe.me]!="factor"]), collapse="+")#get all fixed effects terms together... pot.terms.with.corr=paste(paste("(1+", pot.terms.with.corr, "|", re, ")", sep=""), collapse="+")#... and paste random effects, brackets and all that | Statistical Modeling | https://osf.io/vjeb3/ | diagnostic_fcns.r |
252 | Supplemental Material: To check whether the participants followed gaze at all, we tested the gaze following score against chance level by running a one sample test against zero. Children followed gaze in the additional gazefollowing task (M 3;; SD 1.66;; t(29) 9.89, p .00, d 1.8). Correlation GF & Difference Score | Data$DiffSocCon <- as.numeric(Data$DiffSocCon) cor.test(Data$DS, Data$DiffSocCon, method = "pearson") #only younger | Statistical Test | https://osf.io/4a9b6/ | EXPLORATORY_Gaze Following_ANALYSES.R |
253 | use network data only for students in classrooms with at least 50% response rate | dat$nwinclude <- 0 dat[which(dat$prop_part > .50 & !is.na(dat$prop_part)),"nwinclude"] <- 1 length(unique(dat[which(dat$nwinclude == 1), "IDTESTGROUP_FDZ"])) # 1708 classes with at least 50% response rate length(unique(dat[which(dat$nwinclude == 0), "IDTESTGROUP_FDZ"])) # 290 classes excluded for network related analyses dat[which(dat$nwinclude == 0), c(nw_covariates, grep("deg", names(dat), value = T))] <- NA #recode sociometric variables to NA for classes with low level of completeness length(unique(dat[which(dat$nwinclude == 0), "IDSTUD_FDZ"])) # 5450 students without sociometric data length(unique(dat[which(dat$nwinclude == 1), "IDSTUD_FDZ"])) # 36,920 students with sociometric data | Data Variable | https://osf.io/hu2n8/ | 01_get_gads.R |
254 | Reorder factor levels: (this will be useful once we create the plot legend) | dataset$trial_condition <- factor(dataset$trial_condition, levels = c("diff_3sg", "diff_3pl", "same_3sg", "same_3pl")) dataset$trial_condition_new <- factor(dataset$trial_condition_new, levels = c("different number", "same number")) | Visualization | https://osf.io/37rfb/ | prediction_plots.R |
255 | Prepare dataframe for eyetracking analysis & graphs (package eyetrackingR required) | eyetrackingr.data <- make_eyetrackingr_data(dataset, participant_column = "participant_number", trial_column = "item_nr", time_column = "time_ms_absolute", trackloss_column = "trackloss", aoi_columns = c("average_target_sample_count_proportion", "average_distractor_sample_count_proportion"), treat_non_aoi_looks_as_missing = TRUE ) | Visualization | https://osf.io/37rfb/ | prediction_plots.R |
256 | Remove trackloss per trial > 50% (removed 4 trials) | eyetrackingr.data <- clean_by_trackloss(data = eyetrackingr.data, trial_prop_thresh = .5) | Data Variable | https://osf.io/37rfb/ | prediction_plots.R |
257 | check histogram for binning | hist(quop_use$t1_eff) quop_use$bins<-cut(quop_use$t1_eff,quantile(quop_use$t1_eff,p=seq(0,1,length=29),na.rm=T)) | Visualization | https://osf.io/vphyt/ | Sentence_Level.R |
258 | visualize with forest plot compare corona vs. precorona | forest(lvd$est[1:16],sei=lvd$se[1:16], ilab = cbind(c("Pre-pandemic Cohorts vs. First Pandemic Cohort",rep("",7),"Pre-pandemic Cohorts vs. Second Pandemic Cohort",rep("",7)), paste0("T",rep(1:8,2)), c("Schools Open","???","???","???","Schools Closed","???","???","Alternating Lessons", "Schools Open","???","Alternating Lessons","???","???","???","???","Schools Open")), ilab.xpos = c(-1.35,-.95,-.65), ylim = c(0,22), xlim = c(-1.7,1), slab = NA, rows = c(18:11, 8:1), at = c(-.7,-.35,0,.35,.7), xlab = "Latent Variance Differences", col=c("gray","gray", "gray","gray", "black", "black", "black", "black", "gray", "gray", "black", "black", "black", "black", "black", "gray")) forest(lmd$est[1:16],sei=lmd$se[1:16], ilab = cbind(c("Pre-pandemic Cohorts vs. First Pandemic Cohort",rep("",7),"Pre-pandemic Cohorts vs. Second Pandemic Cohort",rep("",7)), paste0("T",rep(1:8,2)), c("Schools Open","???","???","???","Schools Closed","???","???","Alternating Lessons", "Schools Open","???","Alternating Lessons","???","???","???","???","Schools Open")), ilab.xpos = c(-1.35,-.95,-.65), ylim = c(0,22), xlim = c(-1.7,1), slab = NA, rows = c(18:11, 8:1), at = c(-.7,-.35,0,.35,.7), xlab = "Standardized Latent Mean Differences", col=c("gray","gray", "gray","gray", "black", "black", "black", "black", "gray", "gray", "black", "black", "black", "black", "black", "gray")) | Visualization | https://osf.io/vphyt/ | Sentence_Level.R |
259 | to see how well the distances are measured | stressplot(species_matrix_short.nmds) #R2 = 0.961, linear R2 = 0.891 data.scores <- as.data.frame(scores(species_matrix_short.nmds$points)) data.scores$site <- substr(rownames(data.scores),1,6) data.scores$plots <- substr(rownames(data.scores),8,15) species.scores <- as.data.frame(scores(species_matrix_short.nmds, "species")) species.scores$species <- rownames(species.scores) | Statistical Test | https://osf.io/uq3cv/ | 5_NMDS_comparing_dawn_and_morning_assemblage_compositions.R |
260 | now statistical test to see if communities are statistically different from one another | species_matrix2 <- species_matrix %>% mutate_if(is.numeric, ~1 * (. > 0)) sp_matrix_with_group <- species_matrix2 rownames(sp_matrix_with_group) <- rownames(species_matrix) sp_matrix_with_group$grouping <- substr(rownames(sp_matrix_with_group), 7,13) ano <- anosim(species_matrix2, distance = 'euclidean', grouping = sp_matrix_with_group$grouping) summary(ano) | Statistical Test | https://osf.io/uq3cv/ | 5_NMDS_comparing_dawn_and_morning_assemblage_compositions.R |
261 | SEM Analysis Section The following section sets up a series of path models. Note that this is a highly revised version of the SEM section, which originally tried to incorporate a complex latent variable model to simultaneously test all hypotheses. I decided to break it down into smaller, simpler models, and to focused on observedvariableonly models using the composite meausres instead. I'm not sure if that's the "best" approach or not, but I think it at least makes for a simpler approach. The code below is thus a mixture of what I originally wrote and the revised models. Let's check differences between the two groups: | t.test(Measures$Age~Measures$Normative) t.test(Measures$Adj_Income~Measures$Normative) t.test(Measures$Parent_Edu~Measures$Normative) t.test(Measures$Political~Measures$Normative) t.test(Measures$Crit_Reflection_Mean~Measures$Normative) t.test(Measures$Efficacy_Mean~Measures$Normative) t.test(Measures$Crit_Action_Mean~Measures$Normative) t.test(Measures$BLM_Activism~Measures$Normative) t.test(Measures$Militarism_Mean~Measures$Normative) t.test(Measures$EDNhComp~Measures$Normative) t.test(Measures$ACEs_Sum~Measures$Normative) t.test(Measures$MEQ_Total~Measures$Normative) t.test(Measures$Discrim_Major_Sum~Measures$Normative) t.test(Measures$Discrim_Everyday_Mean~Measures$Normative) t.test(Measures$BNSS_Mean~Measures$Normative) t.test(Measures$BNSSh_Mean~Measures$Normative) t.test(Measures$Zero_Sum_Mean~Measures$Normative) sjmisc::frq(Measures$Race, out = "v") Measures %>% group_by(Race) %>% summarise(BLM_Mean = mean(BLM_Activism)) data <- Measures %>% filter(Race %in% c("Black or African-American","White or Caucasian") & !is.na(BLM_Activism)) %>% mutate(Black = ifelse(Race == "Black or African-American", 1,0)) | Statistical Modeling | https://osf.io/xhrw6/ | 4_sem_models.R |
262 | To assess BLM Activism by race, keep in mind that the BLM measure is count data (with a lot of zeros) A ttest therefore is not appropriate. We will use "zeroinflated Poisson regression" | ggplot(Measures, aes(BLM_Activism)) + geom_histogram() model.zi = zeroinfl(BLM_Activism ~ Black, data = data, dist = "poisson") summary(model.zi) Descriptives3 <- Measures %>% select(Normative,Age,Adj_Income,Parent_Edu,Political,ACEs_Sum,EDNhComp,Discrim_Major_Sum,Discrim_Everyday_Mean,MEXQ_Exp, Crit_Reflection_Mean,Efficacy_Mean,Crit_Action_Mean,BLM_Activism, Militarism_Mean) %>% psych::describeBy(group="Normative") %>% as.data.frame() %>% rownames_to_column(var="Measure") %>% select(-vars) DescrNonNorm<- as.data.frame(Descriptives3[1]) %>% rownames_to_column(var="Measure") DescrNorm<- as.data.frame(Descriptives3[2]) %>% rownames_to_column(var="Measure") | Statistical Modeling | https://osf.io/xhrw6/ | 4_sem_models.R |
263 | Write cleaned covariate files to xlsx | cleaned_group_covariates_data <- list( "cleaned_group_participant" = cleaned_group_participant_dat, "cleaned_group_instructional" = cleaned_group_instructional_dat, "cleaned_group_DV" = cleaned_group_DV_dat, "cleaned_group_race_ethnicity" = race_eth ) write_xlsx(cleaned_group_covariates_data, path = "Cleaned data/Group covariates data tables.xlsx") | Data Variable | https://osf.io/b5ydr/ | 3-Cleancovariatesforgroupdesigns.R |
264 | Weighted regression model for Ordinary People subscale score (mean) | m1.ppl.svyglm.fit <- svyglm(scipopppl ~ age + gender + education.comp + education.uni + sciprox.score + urbanity.log + languageregion.ger + languageregion.ita, design = bar.design.scipopppl, family = gaussian, na.action = na.omit) m2.ppl.svyglm.fit <- svyglm(scipopppl ~ age + gender + education.comp + education.uni + sciprox.score + urbanity.log + languageregion.ger + languageregion.ita + polorientation + religiosity, design = bar.design.scipopppl, family = gaussian, na.action = na.omit) m3.ppl.svyglm.fit <- svyglm(scipopppl ~ age + gender + education.comp + education.uni + sciprox.score + urbanity.log + languageregion.ger + languageregion.ita + polorientation + religiosity + interestscience + sciliteracy + trustscience + trustscientists, design = bar.design.scipopppl, family = gaussian, na.action = na.omit) | Statistical Modeling | https://osf.io/qj4xr/ | 03_explaining-scipop-attitudes.R |
265 | Model assumption checks Do some assumption checks for multiple linear regression (e.g., see Field, 2012, p. 292). These are: (1) Multicollinearity (2) Nonnormality/heteroscedasticity of residuals Assumption checks (1): Multicollinearity Specify model and inspect GVIFs Note: The VIF (usual collinearity diagnostic) may not be applicable to models with dummy regressors constructed from a polytomous categorical variable or polynomial regressors (Fox, 2016: 357). Fox and Monette (1992) introduced generalized variance inflation factor (GVIF) for these cases. As education.uni/education.comp and languageregion.ger/languageregion.ita are such dummy regressors based on polytomous categorial variables (i.e. education and languageregion), we should compute GVIFs. car::vif() does that automatically. GVIFs of SciPop Score models | car::vif(m1.scipopgoertz.svyglm.fit) %>% as.data.frame %>% rename(GVIF = deparse(substitute(.))) car::vif(m2.scipopgoertz.svyglm.fit) %>% as.data.frame %>% rename(GVIF = deparse(substitute(.))) car::vif(m3.scipopgoertz.svyglm.fit) %>% as.data.frame %>% rename(GVIF = deparse(substitute(.))) | Statistical Test | https://osf.io/qj4xr/ | 03_explaining-scipop-attitudes.R |
266 | Nonnormality/heteroscedasticity of residuals of ppl Score models | distributionchecks(m1.ppl.svyglm.fit) distributionchecks(m2.ppl.svyglm.fit) distributionchecks(m3.ppl.svyglm.fit) | Statistical Modeling | https://osf.io/qj4xr/ | 03_explaining-scipop-attitudes.R |
267 | ADDITIONAL ANALYSIS B: TESTING A USHAPED RELATIONSHIP BETWEEN SCIPOP AND POLITICAL ORIENTATION 1st test: Quadratic regression on weighted data | bar2019pop$polorientation2 <- bar2019pop$polorientation^2 bar.design <- svydesign(id = ~0, data = bar2019pop, weights = ~weight) quad.scipopgoertz.svyglm.fit <- svyglm(scipopgoertz ~ age + gender + education.comp + education.uni + sciprox.score + urbanity.log + languageregion.ger + languageregion.ita + polorientation + polorientation2 + religiosity + interestscience + sciliteracy + trustscience + trustscientists, design = bar.design, family = gaussian, na.action = na.omit) summ(model = quad.scipopgoertz.svyglm.fit, binary.inputs = "full", n.sd = 2, transform.response = F, confint = F, vifs = F, scale = F, model.info = T, model.fit = T, digits = 4) quad.scipopgoertz.svyglm.fit$aic m0.scipopgoertz.svyglm.fit <- svyglm(scipopgoertz ~ 1, bar.design, family = gaussian, na.action = na.omit) anova(m0.scipopgoertz.svyglm.fit, quad.scipopgoertz.svyglm.fit, test = "F", method = "Wald") | Statistical Test | https://osf.io/qj4xr/ | 03_explaining-scipop-attitudes.R |
268 | chisquare to compare vegan in experienced vs. not experienced for each advocacy type graphic video | data %>% xtabs(~ graphic_exp_buc + vegn_bin, .) %>% print() %>% proportions("graphic_exp_buc") graphic_tmp <- chisq.test(xtabs(~ graphic_exp_buc + vegn_bin, data)) | Statistical Test | https://osf.io/3aryn/ | 5chisquare_Spanish.R |
269 | Boosted Regression Trees CUS BRT plot | CUS.BRT.plot <- ggplot(data = CUS.BRT, aes(x = predictors, y = influence, fill = predictor.type)) + geom_bar(stat = "identity") + geom_hline(yintercept = 5, linetype = 3, size = 1, colour = "gray60") + scale_fill_manual(values = viridis.3) + coord_flip() + labs(x = "Predictor", y = "Relative Influence") + scale_y_continuous(breaks = c(0, 10, 20, 30, 40), limits = c(0, 40)) + scale_x_discrete(limits = BRT.plot.label.limits, labels = BRT.plot.labels) + theme_pubr() + theme(legend.position = c(0.85, 0.825), legend.title = element_blank()) | Visualization | https://osf.io/62je8/ | DMS-NRSA-CA-QC-Figures.R |
270 | Sets limits of the plot based on user choice | limits <- matrix(0,2,2) limits <- if (lim == 1){ limits <-u.plot.limit(min.age,max.age,lim,min.ylim,max.ylim) } else if (lim == 2){ limits <-u.plot.limit(min.age,max.age,lim,min.ylim,max.ylim) } else { limits <-u.plot.limit(min.xlim,max.xlim,lim,min.ylim,max.ylim) } limconc <-uconcplot(limits[1,1],limits[1,2],int) #sets the concordia line based on plot limts plot_limconc <- limconc[2:(nrow(limconc)-1),] | Visualization | https://osf.io/p46mb/ | U-PbGeochronologyScripts.R |
271 | BELOW CALCULATES THE ELLIPSES calculate x/y coords that define ellipse based in input data including rho | ell.coords=list() #sets up list to write results to for(n in 1:n){ covmat <-cor2cov2((Pb7Ue[n]/2),(Pb6Ue[n]/2),rho[n]) nn <- 75 cutoff <- stats::qchisq(1-0.05,2) e <- eigen(covmat) a <- sqrt(cutoff*abs(e$values[1])) # major axis b <- sqrt(cutoff*abs(e$values[2])) # minor axis v <- e$vectors[,1] beta <- atan(v[2]/v[1]) theta <- seq(0, 2 * pi, length=nn) out <- matrix(0,nrow=nn,ncol=3) out[,1] <- Pb7U[n] + a * cos(theta) * cos(beta) - b * sin(theta) * sin(beta) out[,2] <- Pb6U[n] + a * cos(theta) * sin(beta) + b * sin(theta) * cos(beta) out[,3] <- n | Visualization | https://osf.io/p46mb/ | U-PbGeochronologyScripts.R |
272 | Functions for plotting models Extracts significance from models for chart legend | extractSign <- function (model, modeltype, MtnRanges, InclMIA = TRUE) { coef <- summary(model)$coefficients$cond SignDF <- data.frame(mtnrng = MtnRanges, main = NA, int_ao = NA, int_po = NA) if (InclMIA) { if (modeltype == "int") { if (coef[row.names(coef) == "AO:MtnRange1", 4] < 0.05) {SignDF$int_ao[SignDF$mtnrng == "Coast - North"] <- "I"} if (coef[row.names(coef) == "AO:MtnRange2", 4] < 0.05) {SignDF$int_ao[SignDF$mtnrng == "Coast - South"] <- "I"} if (coef[row.names(coef) == "AO:MtnRange3", 4] < 0.05) {SignDF$int_ao[SignDF$mtnrng == "Columbia - North"] <- "I"} if (coef[row.names(coef) == "AO:MtnRange4", 4] < 0.05) {SignDF$int_ao[SignDF$mtnrng == "Columbia - South"] <- "I"} if (coef[row.names(coef) == "AO:MtnRange5", 4] < 0.05) {SignDF$int_ao[SignDF$mtnrng == "Rockies - North"] <- "I"} if (coef[row.names(coef) == "MtnRange1:PO", 4] < 0.05) {SignDF$int_po[SignDF$mtnrng == "Coast - North"] <- "I"} if (coef[row.names(coef) == "MtnRange2:PO", 4] < 0.05) {SignDF$int_po[SignDF$mtnrng == "Coast - South"] <- "I"} if (coef[row.names(coef) == "MtnRange3:PO", 4] < 0.05) {SignDF$int_po[SignDF$mtnrng == "Columbia - North"] <- "I"} if (coef[row.names(coef) == "MtnRange4:PO", 4] < 0.05) {SignDF$int_po[SignDF$mtnrng == "Columbia - South"] <- "I"} if (coef[row.names(coef) == "MtnRange5:PO", 4] < 0.05) {SignDF$int_po[SignDF$mtnrng == "Rockies - North"] <- "I"} } } SignAOMain <- paste("AO: ", format(round(coef[row.names(coef) == "AO", 1], 3), nsmall=3), " (", format(round(coef[row.names(coef) == "AO", 4], 3), nsmall=3),")") SignPOMain <- paste("PO: ", format(round(coef[row.names(coef) == "PO", 1], 3), nsmall=3), " (", format(round(coef[row.names(coef) == "PO", 4], 3), nsmall=3),")") return(list(SignAOMain = SignAOMain, SignPOMain = SignPOMain, SignDF = SignDF)) } | Visualization | https://osf.io/7xsfj/ | Fig05To08.R |
273 | outlier coding M and SD per subject and item depending on distractor condition and soa (independent variables) | subj.M.SD = ddply(rawdat,.(subj,cond,soa), summarize, subj.M=mean(RT, na.rm=T), subj.SD=sd(RT, na.rm=T)) tar.M.SD = ddply(rawdat,.(targ_ID,cond,soa), summarize, tar.M=mean(RT, na.rm=T), tar.SD=sd(RT, na.rm=T)) rawdat = merge(rawdat, subj.M.SD, by=c("subj", "cond", "soa")) rawdat = merge(rawdat, tar.M.SD, by=c("targ_ID", "cond", "soa")) rawdat$subj.min = (rawdat$subj.M - 2*(rawdat$subj.SD)) rawdat$subj.max = (rawdat$subj.M + 2*(rawdat$subj.SD)) rawdat$tar.min = (rawdat$tar.M - 2*(rawdat$tar.SD)) rawdat$tar.max = (rawdat$tar.M + 2*(rawdat$tar.SD)) | Data Variable | https://osf.io/c93vs/ | exp01_prep.R |
274 | calculate the value of the closest peer estimate and store in 'closest' | closest<-min(abs(si-firstEstimate)) | Data Variable | https://osf.io/rmcuy/ | Fig. |
275 | define Gaussians for updating prior self (firstEstimate) | density_self <- log(dnorm(x, firstEstimate, own_sd )) | Statistical Modeling | https://osf.io/rmcuy/ | Fig. |
276 | 1.1. Sample 1 PTSD Symptoms (PSSI) Run linear mixed effect model | a1_model_01_sg_pssi_outcome <- lme(pssi_end ~ pssi_s0 + sg, random = ~ 1 | id, method = "ML", na.action = na.omit, data = data_a1_pssi) | Statistical Modeling | https://osf.io/dgt8x/ | 04-revisions.R |
277 | Check asssumption of normality of the residuals | qqnorm(resid(a1_model_01_sg_pssi_outcome)) qqnorm(resid(a2_model_01_sg_pssi_outcome)) | Statistical Modeling | https://osf.io/dgt8x/ | 04-revisions.R |
278 | Calculate contrasts specified in in the contrast matrix "contrast_outcome" | a1_model_01_sg_pssi_outcome_contrasts <- glht(a1_model_01_sg_pssi_outcome, contrast_outcome) a2_model_01_sg_pssi_outcome_contrasts <- glht(a2_model_01_sg_pssi_outcome, contrast_outcome) | Statistical Modeling | https://osf.io/dgt8x/ | 04-revisions.R |
279 | Show contrasts without adjusting pvalues for multiple comparisons | summary(a1_model_01_sg_pssi_outcome_contrasts, test = adjusted("none")) summary(a2_model_01_sg_pssi_outcome_contrasts, test = adjusted("none")) | Statistical Test | https://osf.io/dgt8x/ | 04-revisions.R |
280 | Select estimates, standard error and pvalue | table_a1_model_01_sg_pssi_outcome_full <- table_a1_model_01_sg_pssi_outcome_raw %>% dplyr::select(beta = estimate, se = std.error, p = p.value) table_a2_model_01_sg_pssi_outcome_full <- table_a2_model_01_sg_pssi_outcome_raw %>% dplyr::select(beta = estimate, se = std.error, p = p.value) | Data Variable | https://osf.io/dgt8x/ | 04-revisions.R |
281 | Calculate pooled standard deviation at baseline | sd_a1_pssi_s0 <- sd(data_a1_pssi$pssi_s0, na.rm = TRUE) sd_a2_pssi_s0 <- sd(data_a2_pssi$pssi_s0, na.rm = TRUE) | Statistical Modeling | https://osf.io/dgt8x/ | 04-revisions.R |
282 | Calculate Cohen's d using the absolute value of the baseline adjusted difference from the linear mixed effect model | cohens_d_a1_pssi_end <- a1_pssi_end_diff / sd_a1_pssi_s0 cohens_d_a2_pssi_end <- a2_pssi_end_diff / sd_a2_pssi_s0 | Statistical Test | https://osf.io/dgt8x/ | 04-revisions.R |
283 | Peer Rank Means and Standard Deviations of Conditions Unspecified Peer Rank | round(mean(unspecified$peer.rank), 2) round(sd(unspecified$peer.rank), 2) | Data Variable | https://osf.io/9tnmv/ | Exp4_buddhist_post.R |
284 | determines and rounds (to digits as specified in digits) mean and sd of x, and then collapses them in a single entry | xx=as.character(round(c(mean(x, na.rm=T), sd(x, na.rm=T)), digits=digits)) if(any(grepl(x=xx, pattern=".", fixed=T))>0){ xx[!grepl(x=xx, pattern=".", fixed=T)]=paste(xx[!grepl(x=xx, pattern=".", fixed=T)], "0", sep=".") xx=matrix(unlist(strsplit(as.character(xx), split=".", fixed=T)), ncol=2, byrow=T) xx[, 2]=unlist(lapply(xx[, 2], function(x){ paste(c(x, paste(c(rep("0", times=digits-nchar(x))), collapse="")), collapse="") })) xx=apply(xx, 1, paste, collapse=".") } paste(xx, collapse=sep) } c.tab<-function(x, digits=NA, n.spaces=1, add.hash=T, incl.fst=F, incl.rownames=T){ | Data Variable | https://osf.io/vjeb3/ | helpers.r |
285 | wrapper for savePlot with file "clipboard" and type "wmf" (default;; others are possible) | savePlot(file="clipboard", type=type) } overdisp.correction<-function(coeffs, disp.param){ coeffs[, "Std. Error"]=coeffs[, "Std. Error"]*sqrt(disp.param) coeffs[, "z value"]=coeffs[, "Estimate"]/coeffs[, "Std. Error"] coeffs[, "Pr(>|z|)"]=2*pnorm(q=-abs(coeffs[, "z value"]), mean=0, sd=1) return(coeffs) } merge.ests.ci.stab<-function(coeffs, ci=NULL, tests=NULL, stab=NULL){ ires=coeffs if(!is.null(ci)){ coeffs=cbind(coeffs, ci[rownames(coeffs), ]) } if(!is.null(tests)){ xx=outer(rownames(tests), rownames(coeffs), Vectorize(function(tt, e){ if(nchar(tt)>nchar(e)){ return(0) }else{ | Visualization | https://osf.io/vjeb3/ | helpers.r |
286 | Add the starttime (intra_scope_window[1]) to every start position (scope$start) ... ... of the recording timestamp to get the actual start window in milliseconds | starting_times <- df$RecordingTimestamp[scope$start] + as.numeric(intra_scope_window[1]) } if (intra_scope_window[2] != "end") { | Data Variable | https://osf.io/mp9td/ | get_looks.R |
287 | set to current trial duration to 0 | current_trial_total_duration[[hn]] <- 0 } current_trial_total_looks[[hn]] <- 0 } current_first_look_duration[[hn]] <- 0 current_first_look_ending_reason[[hn]] <- "" first_looks_collection[[hn]]$found_first <- FALSE first_looks_collection[[hn]]$forced_stop <- FALSE } | Data Variable | https://osf.io/mp9td/ | get_looks.R |
288 | Return sum log likelihood. Include protection against log(0) problems | return(sum(log(pmax(like, 1e-10)))) } else { | Statistical Modeling | https://osf.io/tbczv/ | exp1bDead-MNL-SAT-A.r |
289 | Plot posterior fit and forward simulation prediction | plot_fits <- function(stan_fit_ex, N_t, N_p, obs_times, pred_times, data_vector, S_0, D_0, M_0, filename) { sigma_hat <- median(stan_fit_ex$sigma) CO2_flux_ratios_hat_median <- rep(NA, N_t) #Pre-allocate vector for median model fit CO2_flux_ratios_pred_median <- rep(NA, N_p) sigma_hat <- median(stan_fit_ex$sigma) for (t in 1:N_t) { CO2_flux_ratios_hat_median[t] <- median(stan_fit_ex$CO2_flux_ratios_hat_vector[,t]) } for (t in 1:N_p) { CO2_flux_ratios_pred_median[t] <- median(stan_fit_ex$CO2_flux_ratios_new_vector[,t]) } | Visualization | https://osf.io/7mey8/ | stan_CON_adriana_pools5i.r |
290 | fit linear mixed effects models null model | fitlmer0 <- lmer(rating ~ 1 + (1|subject), data = dat.long1, REML = FALSE) summary(fitlmer0) | Statistical Modeling | https://osf.io/eg6w5/ | experiment1c_analyses.R |
291 | the mean of each parameter across iterations. Keep dimensions for parameters and subjects | mean.params <- t(apply(sampled$samples$alpha[,,keep],1:2,mean)) | Statistical Modeling | https://osf.io/wbyj7/ | pmwg-DIC.r |
292 | plot the calibration curve (Abs vs. [BSA]) with an appropriate trendline and figure caption | ggplot(bca, aes(x = BSA, y = Abs562)) + | Visualization | https://osf.io/9e3cu/ | BCA_activity_answers.R |
293 | generate MPT data: | gendat <- genMPT(theta = pnorm(theta), restrictions = "model/restrictions.txt", numItems = numItems, eqnfile="model/2htsm.eqn") | Data Variable | https://osf.io/s82bw/ | 04_simulation_continuous_predictor.R |
294 | Recode confirm(oppose/confirm) and won(win/lost) to factors | DF$confirm <- recode(DF$confirm, `0` = "Confirmed", `1` = "Opposed") DF$won <- recode(DF$won, `0` = "Lost", `1` = "Won") DF <- DF %>% mutate(confirm = as_factor(confirm)) %>% mutate(treatment = as_factor(treatment)) %>% mutate(won = as_factor(won)) %>% mutate(majmin = as_factor(majmin)) %>% mutate(experiment = as_factor(experiment)) %>% mutate(evidence = as_factor(evidence)) summary(DF) | Data Variable | https://osf.io/9gjyc/ | BRMS + Figure4.R |
295 | create unique id number for each unique individual | mutate(uniq.id = group_indices(., id, experiment)) %>% mutate(uniq.id = as_factor(uniq.id)) summary(Ind_E1) M_Ind_E1 <- brm(changed ~ won*confirm + trial + (1|uniq.id), data = Ind_E1, family = "bernoulli", iter = 6000, chains = 3, cores = 3, save_all_pars = TRUE, file = "Ind_E1") summary(M_Ind_E1) plot(M_Ind_E1, ask = FALSE) pp_check(M_Ind_E1, check = "distributions") pp_check(M_Ind_E1, check = "residuals") pp_check(M_Ind_E1, "error_scatter_avg") pp_check(M_Ind_E1, check = "scatter") | Data Variable | https://osf.io/9gjyc/ | BRMS + Figure4.R |
296 | generates 95% confidence intervals for each beta coefficient | m.p1_decision.CI <- round(confint(m.p1_decision, parm = "beta_"), 3) | Statistical Test | https://osf.io/uygpq/ | Within-paradigm.R |
297 | exclude people who don't meet the inclusion criteria of having data for at least 3 days in each week | person_weekly_entries <- dat2019 %>% group_by(identity_id, woy) %>% summarise(n_days = length(identity_id)) person_weekly_entries <- as.data.frame(person_weekly_entries) person_weekly_entries <- subset(person_weekly_entries, n_days >= 3) | Data Variable | https://osf.io/wyrav/ | Dataprep-ParticipantResponses.R |
298 | construct a maximal glmer() model This model contains a fixed withinsubjects effect of Ambiguity (effectcoded with 0.5 amb), and betweensubjects fixed effects for Vocabulary and ART test scores, plus random effects by participants and items. | Acc.max <- glmer(accuracy ~ 1 + Ambiguity.code + Vocab.Cent + ART.Cent + Ambiguity.code : Vocab.Cent + Ambiguity.code : ART.Cent + (1 + Ambiguity.code | RECORDING_SESSION_LABEL) + (1 + Vocab.Cent + ART.Cent | item), data = Data.CohOnly, family = "binomial", control = glmerControl(optimizer ="bobyqa")) | Statistical Modeling | https://osf.io/hn3bu/ | AnalysisCode.R |
299 | construct a maximal lmer() model This model contains a fixed withinsubjects effect of Ambiguity (effectcoded with 0.5 amb), and betweensubjects fixed effects for Vocabulary and ART test scores, plus random effects by participants and items. Maximal model with interactions | TrialDwellTime.max <- lmer(logTrialDwellTime ~ 1 + Ambiguity.code + Vocab.Cent + ART.Cent + Ambiguity.code : Vocab.Cent + Ambiguity.code : ART.Cent + (1 + Ambiguity.code | RECORDING_SESSION_LABEL) + (1 + Vocab.Cent + ART.Cent | item), data = Data.CorrTrials, REML=FALSE) | Statistical Modeling | https://osf.io/hn3bu/ | AnalysisCode.R |
300 | _firstfixation duration Construct a maximal lmer() model Because not all the necessary models converge, remove correlations between random effects for all models for the firstfixation duration measure: | AOIKey.FF.reduced <- lmer(logFirstFixDur ~ 1 + Ambiguity.code + Vocab.Cent + ART.Cent + Ambiguity.code : Vocab.Cent + Ambiguity.code : ART.Cent + (1 | RECORDING_SESSION_LABEL) + (0 + Ambiguity.code | RECORDING_SESSION_LABEL) + (1 | item) + (0 + Vocab.Cent | item) + (0 + ART.Cent | item), data = Data.CorrTrials, REML=FALSE) AOICohcue.FF.reduced <- lmer(logFirstFixDur ~ 1 + Ambiguity.code + Vocab.Cent + ART.Cent + Ambiguity.code : Vocab.Cent + Ambiguity.code : ART.Cent + (1 | RECORDING_SESSION_LABEL) + (0 + Ambiguity.code | RECORDING_SESSION_LABEL) + (1 | item) + (0 + Vocab.Cent | item) + (0 + ART.Cent | item), data = Data.CorrTrials, REML=FALSE) | Statistical Modeling | https://osf.io/hn3bu/ | AnalysisCode.R |
Subsets and Splits