Preregistration

We preregistered this study with the Center for Open Science’s Open Science Framework.

  OSF preregistered pre-analysis plan »


Study information

Title

Procrastination and self-rated health: A meta-analysis of own data (MOD)

Authorship

  • Fuschia Sirois
  • Cormac Monaghan

Research Questions

Note  OSF question

Please list each research question included in this study

This study meta-analyses the association between trait procrastination and self-rated health (SRH), with a focus on identifying potential mediators and moderators of this relationship. The following four research questions guide the investigation:

  1. Is trait procrastination associated with lower levels of current and future self-rated health?
  2. Does perceived stress account for the relationship between trait procrastination and both current and future self-rated health?
  3. Do sample characteristics or measurement differences moderate the relationship between procrastination and self-rated health / procrastination and future self-rated health, before and after accounting for the contributions of stress to these associations?
  4. Do demographic factors such as age and sex moderate the relationship procrastination and self-rated health / procrastination and future self-rated health, before and after accounting for stress?

Hypotheses

Note  OSF question

For each of the research questions listed in the previous section, provide one or multiple specific and testable hypotheses. Please state if the hypotheses are directional or non-directional. If directional, state the direction. A predicted effect is also appropriate here.

Primary associations

  1. Hypothesis 1a (Directional): Trait procrastination will be negatively associated with self-health (SRH).
  2. Hypothesis 1b (Directional): Trait procrastination will also be negatively associated with future self-rated health.

Rationale: Prior research suggests that procrastination is linked to poorer health outcomes (Sirois, 2015, 2023; Sirois & Biskas, 2024) and SRH is a broad indicator of well-being (Jylhä, 2009).

Associations (controlling for stress)

  1. Hypothesis 2a (Directional): Trait procrastination will make a unique, negative contribution to self-rated health after statistically controlling for stress - this adjusted effect will be smaller in magnitude than the unadjusted correlation.
  2. Hypothesis 2b (Directional): Trait procrastination will make a unique, negative contribution future self-rated health after statistically controlling for stress - this adjusted effect will similarly be smaller than the unadjusted correlation

Rationale: The Procrastination Health Model (Sirois et al., 2003; Sirois, 2007) suggests stress partially mediates the relationship between procrastination and health, thus attenuating the direct association when controlled.

Moderation by sample and measurement characteristics

  1. Hypothesis 3 (Non-Directional): The unadjusted and adjusted (semi-partial) correlations between trait procrastination and both current and future self-rated health will not significantly differ as a function of sample type (e.g., student, or community) or the procrastination measure

Rationale: While variability in sample and measurement tools could influence effect sizes, the hypotheses assume no systematic moderation.

Moderation by demographic characteristics

  1. Hypothesis 4 (Non-Directional): The unadjusted and adjusted (semi-partial) correlations between trait procrastination and self-rated health will not significantly differ by participant age or sex.

Rationale: Although demographic differences may impact health outcomes generally, their influence on the procrastination - SRH link is not expected to be consistent or significant across datasets

Data Description

Datasets used

Note  OSF question

Name and briefly describe the dataset(s), and if applicable, the subsets of the data you plan to use. Useful information to include here is the type of data (e.g., cross-sectional or longitudinal), the general content of the questions, and some details about the respondents. In the case of longitudinal data, information about the survey’s waves is useful as well. Mention the most relevant information so that readers do not have to search for the information themselves.

This project involves the analysis of 36 datasets collected from the project team’s labs. Of these datasets, 7 were sampled from university student populations, 27 were sampled from general adult and community populations, and the remaining 2 were sampled from medical populations. Additionally, 21 of these samples contain a measure of future self-rated health, 13 with a measure of stress, of these samples, 7 contain a measure of both future self-rated health and stress within the same sample.

Data access

Note  OSF question

If there are any restrictions to accessing the dataset, please describe this here.

Ethical permission for sharing with other researchers was obtained for only 10 of the 33 data sets during the consent process. For all other data sets, ethical permission to share the data outside the research team was not obtained at the time of collection, which was prior to when Open Science practices were routine. As such these data sets cannot be shared.

Access date

Note  OSF question

Specify the download or data access date. If the data were accessed multiple times by different team members, specify the download date for that data that will be used in the statistical analysis.

Not applicable as the data analysed is secondary data from the researchers’ own labs.

Data collection procedures

Note  OSF question

If the data collection procedure is well documented, provide a link to that information. If the data collection procedure is not well documented, describe, to the best of your ability, how data were collected. Describe the representativeness of the sample and any possible biases stemming from the data collection.

Of the 33 data sets, all data were collected online except for samples 4, where 264 of the responses were completed via a mail in survey. All but one of the data sets were convenience samples collected from the project team’s labs. Sample 4 was recruited from a list of 1,000 participants randomly sampled from a nursing association.

Codebook

Note  OSF question

Some studies offer codebooks to describe their data. If such a codebook is publicly available, link, cite, or upload the document. If not, provide other available documentation. Also provide guidance on what parts of the codebook or other documentation are most relevant.

As this current project involves 33 data sets, no single codebook is available. Instead, the list of measured variables summarises the variable names for the main variables analysed which have been coded consistently across the data sets.

Variables

Measured variables

Note  OSF question

Describe both outcome measures as well as predictors and covariates and label them accordingly. If you are using a scale or an index, state the construct the scale/index represents, which items the scale/index will consist of, and how these items will be aggregated. When the aggregation is based on exploratory factor analysis (EFA) or confirmatory factor analysis (CFA), also specify the relevant details (EFA: rotation, how the number of factors will be determined, how best fit will be selected, CFA: how loadings will be specified, how fit will be assessed, which residuals variance terms will be correlated). If you are using any categorical variables, state how you will code them in the statistical analyses.

Demographic variables as moderators

  • Age (numeric value)
  • Sex (1 = female; 2 = male)
    • These values may be reversed for some data sets and will be indicated as such in each data file.
    • Percentage of sample who were female was used as a continuous moderator across the samples.

Main analysis variables

All data sets included measures of:

  • Trait procrastination
  • Self-rated health (both current self-rated health and future self-rated health)
  • A measure of perceived stress

All scales are well validated with their details described elsewhere in the research literature.

Outcome variable

  • Self-rated health was measured using a single item measure asking respondents to rate their physical health on a scale to 1 (excellent) to 5 (poor).
  • Future self-rated health was measured using a single item measure asking respondents to reflect upon and rate their physical health in 10 years times on a scale of 1 (excellent) to 5 (poor).

Predictor variables

  • Trait procrastination was measured with the long 20-item (Lay, 1986) or short 9-item (Sirois et al., 2019), general procrastination scale (GPS).
    • Items from the GPS are score on a 5-point Likert type scale.
  • Perceived stress was measured using the 10 item Perceived Stress Scale (PSS;
    • Items from the PPS are scored on a 5-point Likert type scale.

Missing data

Note  OSF question

What do you know about missing data in the dataset (i.e., overall missingness rate, information about differential dropout)? How will you deal with incomplete or missing data? Provide descriptive information, if available, on the amount of missing data for each variable you will use in the statistical analyses. Based on this information, provide a new expected sample size.

  • As this analysis will include both a re-analysis of some previously analysed data sets, and new analysis of data not previously analysed with respect to the variables of interest, it is not possible to estimate the overall missingness rate.
  • Only cases with less than 20% missing data on each of the key variables will be included.
  • Once the data is tested for missingness using Little’s MCAR test as a guide, any missing data will be replaced via linear interpolation in the case that the MCAR test is non-significant, or multiple imputation (multiple imputation using chained equations or full information maximum likelihood) if the MCAR test is significant and other guidelines are considered.
  • We will follow the guidelines from Jakobsen et al. (2017) to determine if multiple imputation is needed to replace missing data

Units of analysis

Note  OSF question

Which units of analysis (respondents, cases, etc.) will be included or excluded in your study? Taking these inclusion and exclusion criteria into account, indicate the expected sample size of the data you’ll be using for your statistical analyses. If you have a research question about a certain group you may need to exclude participants based on one or more characteristics. Be very specific when describing these characteristics so that readers will be able to redo your moves easily.

All useable data for respondents 18 years and older will be included in the analysis.

Statistical outliers

Note  OSF question

How will you define what a statistical outlier is in your data and what will you do when you encounter them? If you plan to remove outliers, provide a new expected sample size. If you expect to remove many outliers or if you are unsure about your outlier handling strategy, it is good practice to preregister analyses including and excluding outliers. Note that this will be the definitive expected sample size for your study and you will use this number to do any power analyses.

  • No outliers will be removed from the individual data sets.
  • Sensitivity analysis will be conducted for any data sets that produce results which are vastly different from the other data sets
    • Correlations that are in the opposite direction to the other data sets and what is expected or are of a magnitude that is noticeably different from the that from the other data sets
  • Sensitivity analysis will also be conducted for any data sets that measure the key variables in ways that deviate from the standard scaling.

Knowledge of Data

Prior publication/Dissemination

Note  OSF question

List the publications, working papers, and conference presentations you have worked on that are based on the dataset you will use. For each work, list the variables you analysed, but limit yourself to variables that are relevant to the proposed analysis. If the dataset is longitudinal, also state which wave of the dataset you analysed. Specify the previous works for each co-author separately. A list of all papers that have been published from the data sets is included below. Only data from the first time point will be used for the analysis to ensure consistency with the predominantly cross-sectional data.

Papers published from the data sets and variables analysed that overlap with the current study protocol:

  • Sirois, F. M., & Tosti, N. (2012). Lost in the moment? An investigation of procrastination, mindfulness, and well-being. Journal of Rational-Emotive & Cognitive-Behavior Therapy, 30 (4), 237-248.
    • Trait procrastination
    • Perceived stress
  • Sirois, F. M., Molnar, D. M., & Hirsch, J. K. (2017). A meta-analytic and conceptual update on the associations between procrastination and multidimensional perfectionism. European Journal of Personality, 31, 137-159.
    • Trait procrastination
  • Sirois, F. M., & Molnar, D. S. (2017). Perfectionism strivings and concerns are differentially associated with self-rated health beyond negative affect. Journal of Research in Personality, 70, 73-83.
    • Self-rated health
  • Sirois, F. M., & Biskas, M. (2024). Procrastination and health in nurses: Investigating the roles of stress, health behaviours and social support. International Journal of Environmental Research and Public Health, 21, 898
    • Trait procrastination
    • Perceived stress
    • Self-rated health
  • Development and validation of the GPS-9, a short and reliable measure of trait procrastination. Personality and Individual Differences.146, 26-34
    • Trait procrastination
    • Perceived stress
  • Sirois, F. M. (2020). The association between self-compassion and self-rated health in 26 samples. BMC Public Health, 20:74
    • Self-rated health
  • Sirois, F. M., Nauts, S., & Molnar D. S. (2019). Self-compassion and bedtime procrastination: An emotion regulation perspective. Mindfulness, 10, 434–445.
    • No variables analysed
  • Sirois, F. M. (2014). Procrastination and stress: Exploring the role of self-compassion. Self and Identity, 13 (2), 128-145.
    • Trait procrastination
    • Stress
  • Sirois, F. M., *Kitner, R., & Hirsch, J. K. (2015). Self-compassion, affect, and health-promoting behaviors. Health Psychology, 34, 661-669.
    • No variables analysed
  • Sirois, F. M., & *Kitner, R. (2015). Less adaptive or more maladaptive? A meta-analytic investigation of procrastination and coping. European Journal of Personality, 29, 434-444
    • Trait procrastination
  • Sirois, F. M. (2014). Absorbed in the moment? An investigation of procrastination, absorption and cognitive failures. Personality and Individual Differences, 71, 30-34.
    • Trait procrastination
  • Sirois, F. M., van Eerde, W., & Argiropoulou, M. I. (2015). Is procrastination related to sleep quality? Testing an application of the procrastination-health model. Cogent Psychology, 2 (1).
    • Trait procrastination

Prior knowledge

Note  OSF question

Disclose any prior knowledge you may have about the dataset that is relevant for the proposed analysis. If you do not have any prior knowledge of it, please state so. Your prior knowledge could stem from working with the data first-hand, from reading previously published research, or from codebooks. Provide prior knowledge for every author separately.

As all data sets are from the authors’ labs, we necessarily have previous knowledge of the data sets that originate from our own lab.

Analysis Plan

Statistical models

Note  OSF question

For each hypothesis, describe the statistical model you will use to test the hypothesis. Include the type of model (e.g., ANOVA, multiple regression, SEM) and the specification of the model. Specify any interactions and post-hoc analyses and remember that any test not included here must be labelled as an exploratory test in the final paper.

Data will be analysed using both a combination of R and SPSS. Pearson’s correlations will be calculated for the associations of trait procrastination and self-rated health, and for each of the associations of stress with trait procrastination and self-rated health in each sample that measured these variables.

Following this, we will calculate the semi-partial correlations of trait procrastination and self-rated health adjusting individually for the contribution of perceived stress. Stress will be partialled out of the IV (trait procrastination). This will yield one set of adjusted effects in addition to the bivariate correlations to be meta-analysed.

A random effects meta-analysis will be conducted on the calculated effects using the Comprehensive Meta-Analysis (CMA; Version 4) software. CMA first transforms the individual correlation coefficients into Fisher’s Z scores before meta-analysing the effects. When there is more than one time point, the effects at baseline will be used.

Moderator analyses will only be conducted for effects that are significant, and for which there is significant heterogeneity. The degree of variability among the pool of effect sizes, i.e., correlations and semi-partial correlations of trait procrastination and self-rated health, and those that partial out the contribution of stress from trait procrastination, will be assessed with the heterogeneity statistic, \(Q\), to determine if moderator analyses are warranted. The \(I^2\) statistic will also be examined as an additional test of heterogeneity to estimate the proportion of variability present that is not due to sampling error within studies. If the \(I^2\) statistic is associated with a large confidence interval, and the \(I^2\) values are \(> 50\%\) (reflecting moderate heterogeneity), then moderator analyses will be conducted if viable.

Categorical moderators (i.e., sample type, measure type) will be analysed using subgroup analysis with a mixed effects approach where the combined subgroups are first analysed with a random effects model to further assess heterogeneity within each subgroup and then combined using a fixed effects model to assess the heterogeneity between subgroups. Individual levels of heterogeneity with each subgroup will also be reported

Continuous moderators, such as age and percent female in the sample, will be analysed using meta-regression.

Effect size

Note  OSF question

If applicable, specify a predicted effect size or a minimum effect size of interest for all the effects tested in your statistical analyses.

Not applicable as we will be meta-analysing the effects sizes across 33 samples.

Inference criteria

Note  OSF question

What criteria will you use to make inferences? Describe the information you will use (e.g. specify the p-values, effect sizes, confidence intervals, Bayes factors, specific model fit indices), as well as cut-off criteria, where appropriate. Will you be using one-or two-tailed tests for each of your analyses? If you are comparing multiple conditions or testing multiple hypotheses, will you account for this, and if so, how?

We will make inferences about the associations of trait procrastination and self-rated health (current and future), and those after accounting for the contribution of stress, based on the p values and confidence intervals for each of the meta-analyses of the effects across the samples.

  • We will conclude that the average effect size garnered from each analysis supports our hypotheses if \(p < 0.05\) and the confidence interval for each average effect does not include 0
  • For the moderator analyses we will conclude that moderators are not significant if \(p > 0.05\) and the confidence interval for each average effect includes 0.
  • Following established guidelines (Cohen, 1988), the magnitudes of the effect sizes will be evaluated as followed:
    • \(r = 0.10\) considered a small effect size
    • \(r = 0.30\) considered a medium effect size
    • \(r = 0.50\) considered a large effect size

Assumption violation/Model non-convergence

Note  OSF question

What will you do should your data violate assumptions, your model not converge, or some other analytic problem arises?

  • For continuous moderators, a minimum of 10 samples is necessary for a meta-regression, as suggested by the Cochrane collaboration.
  • Subgroup moderator analyses will only be conducted if there are three or more studies in each subgroup.

Reliability and robustness resting

Note  OSF question

Provide a series of decisions or tests about evaluating the strength, reliability, or robustness of your finding. This may include within-study replication attempts, additional covariates, cross-validation, applying weights, selectively applying constraints in an SEM context (e.g., comparing model fit statistics), overfitting adjustment techniques used, or some other simulation/sampling/bootstrapping method.

Overall, the analytic approach we are taking involves replicating the effects across multiple samples (in our case 33 samples) and therefore provides a test of robustness of the effects. We will additionally undertake the steps below to provide further tests of reliability and robustness of the effects.

  • We will conduct a sensitivity analysis for any data sets that include measures of the key variables that are not consistent with the overall pool of data sets, by removing the sample(s) with the odd measure and re-meta-analyse the data to see if the average effect is comparable or different from that including all samples. If it is not, we will conclude that the findings are robust to the inclusion of the sample with the different measure.
  • To estimate of the number of studies with null results that would have to be included in the meta-analysis to render the current findings non-significant, we will calculate a Failsafe N. Rosenthal (1979) guidelines were followed for determining an adequately high fail-safe N. As such, we calculate N as: \(N = 5k + 10\), where k represents the number of studies included.

Exploratory analysis

Note  OSF question

If you plan to explore your dataset to look for unexpected differences or relationships, describe those tests here. If reported, add them to the final paper under a heading that clearly differentiates this exploratory part of your study from the confirmatory part.

In addition to the confirmatory hypotheses outlined above, we plan to conduct several exploratory analyses to investigate patterns not captured by the primary research questions.

  1. First, if the overall (average) effect of procrastination on self-rated health is not significant, we may explore the potential role of unanticipated moderators by conducting additional moderation analyses. These may include variables not initially specified (e.g., education level, cultural background) that emerge as potentially influential during the initial data review.
  2. Second, we may conduct sensitivity analyses in cases where findings from one or more datasets deviate substantially from the overall pattern. This may include examining variation in effect sizes, measurement approaches (e.g., different scales), or distinct sample characteristics that could influence results.
  3. Finally, we are interested in exploring the discrepancy between current and future self-rated health. Specifically, we will examine whether trait procrastination is associated with a larger gap between participants’ current self-rated health and their anticipated future health. This analysis is exploratory and aims to provide initial insights into how temporal health expectations may relate to self-regulatory traits such as procrastination.
  4. All exploratory results will be clearly reported in a separate section of the final manuscript and interpreted cautiously, without the same level of inferential weight as the pre-specified confirmatory analyses.

References

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed). L. Erlbaum Associates.
Jakobsen, J. C., Gluud, C., Wetterslev, J., & Winkel, P. (2017). When and how should multiple imputation be used for handling missing data in randomised clinical trials–a practical guide with flowcharts. BMC Medical Research Methodology, 17(1), 162. https://doi.org/10.1186/s12874-017-0442-1
Jylhä, M. (2009). What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social Science & Medicine, 69(3), 307–316. https://doi.org/10.1016/j.socscimed.2009.05.013
Lay, C. H. (1986). At last, my research article on procrastination. Journal of Research in Personality, 20(4), 474–495. https://doi.org/10.1016/0092-6566(86)90127-3
Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86(3), 638. https://doi.org/10.1037/0033-2909.86.3.638
Sirois, F. M. (2007). Ill look after my health, later: A replication and extension of the procrastinationhealth model with community-dwelling adults. Personality and Individual Differences, 43(1), 15–26. https://doi.org/10.1016/j.paid.2006.11.003
Sirois, F. M. (2015). Is procrastination a vulnerability factor for hypertension and cardiovascular disease? Testing an extension of the procrastination–health model. Journal of Behavioral Medicine, 38(3), 578–589. https://doi.org/10.1007/s10865-015-9629-2
Sirois, F. M. (2023). Procrastination and stress: A conceptual review of why context matters. International Journal of Environmental Research and Public Health, 20(6), 5031. https://doi.org/10.3390/ijerph20065031
Sirois, F. M., & Biskas, M. (2024). Procrastination and Health in Nurses: Investigating the Roles of Stress, Health Behaviours and Social Support. International Journal of Environmental Research and Public Health, 21(7), 898. https://doi.org/10.3390/ijerph21070898
Sirois, F. M., Melia-Gordon, M. L., & Pychyl, T. A. (2003). I’ll look after my health, later: an investigation of procrastination and health. Personality and Individual Differences, 35(5), 1167–1184. https://doi.org/10.1016/S0191-8869(02)00326-4
Sirois, F. M., Yang, S., & Eerde, W. van. (2019). Development and validation of the general procrastination scale (GPS-9): A short and reliable measure of trait procrastination. Personality and Individual Differences, 146, 26–33. https://doi.org/10.1016/j.paid.2019.03.039