I have a *free* textbook called How2statsbook: www.how2statsbook.com. You can download the chapters (1 to 16) as PDF files.
The textbook has approximately 400 RU-vid videos that are hosted on my other RU-vid channel: How2statsbook:
ru-vid.com/show-UClow2dWW7y1uO4kK4SxwsjA
More chapters to come.
How2stats is dedicated to helping people learn about statistics. Some of the videos are more lecture based, while others are relevant to using software for the purposes of conducting statistical analyses.
I try to cover topics in a non-mathematical way, so that nearly anyone can understand. I usually try to go the extra mile, so that people can truly understand what the statistical analysis is doing and how it compares to alternative statistical procedures.
Finally, I also point out some statistical mistakes that people often make, as well as some tips and tricks to help people get the best results.
I have an MS in statistics and broadly agree with you. There are times when it does make sense to apply some kind of correction, but in most cases, it probably makes better sense to publish the results and simply talk about it as a limitation or area of further exploration in the discussion, like you said, especially if the p value is just under 0.05. I would argue we need to collectively move away from viewing the 0.05 as a black or white judgement on whether research in meaningful or not. Even an p value of 0.10 *might* indicate there is a relationship there that would show up more significantly with a larger sample size or if some other factor of the experiment was controlled for. Also, thank you for the references. I will be using them.
splitting continuous data into categorical data (low, mod, high) is not generally recommended. Is it possible that the same score is divided in mod and high? (couldn't see that one in your dataset but it would be plausible and less than ideal) However I do like the graphic nature of the regression lines and will think on this method some more for moderation
I cant’t thank you enough for these videos, I would be working on my assignment for hours more without these, I love your explanations and I genuinely was learning. Thanks again!
Thank you for so adequately explaining the differences in descriptive statistics from one part of the analysis to another. This confused me sooooo much and was not covered in my stats module!!!
Thank you so much. Could you please make a video expaining how to use ANCOVA in 2x2 within subjects ANOVA? You have one video for ANCOVA related to 2x2 between subjects ANOVA.
@how2stats What are the mean achievement scores in male and female Students taught social studies using digital instructional material and non digital instructional material at pretest and post test What about this?
I like it, but have you ever considered using Latex? The format would improve a lot (I am not saying it's not good, only that it is worth to consider using Latex instead). Warm regards! :)
How about this syntax to compare against another sample’s t values (but with mean and SD known): data list list / n1 sample_mean1 sample_mean2 sample_sd1 sample_sd2 n2. begin data 41 136 124 19 16 185 end data. Compute mean_difference = sample_mean1 - sample_mean2. Compute square_root_n1 =SQRT(n1). Compute square_root_n2 =SQRT(n2). Compute pooled_sd =SQRT((sample_sd1*sample_sd1 + sample_sd2*sample_sd2)/2). Compute standard_difference = pooled_sd/(square_root_n1 + square_root_n2). Compute t_statistic = mean_difference/standard_difference. Compute p_value = 2*(1 - CDF.T(N1 + N2 - 2, ABS(t_statistic))). Compute cohens_d = mean_difference/pooled_sd. EXECUTE. Formats t_statistic p_value cohens_d (f8.5). LIST t_statistic p_value cohens_d. Obtained through ChatGPT (I had to edit a lot of rubbish out to fix errors) by putting in your syntax and asking it to change it for sample to sample comparison t statistic. Still getting error for thep value
Can you please tell me what happens if I had 55 students from a class that I taught, I took pre-test scores, but only 35 participated, then I did 2 months of teaching and now 40 participated in post-test, and let's say that only 30 are the same (pre and post), can I assume that they came from the same population so I can use all data or should I only use data where their initials filled out on the forms are the same (exclude those who weren't there in pre-test) and have a much smaller sample? And if I can assume the first; they came from the same population (they were in the same class), should I use paired or unpaired sample T-test considering some who are in post, didn't fill out pre (and vice versa)?
Here's an expanded summary of the video on performing a one-way ANOVA in SPSS, including more detailed insights: Introduction to One-Way ANOVA: The video begins with an introduction to one-way ANOVA, emphasizing its use in comparing means between groups. This analysis helps determine if there are significant differences among group averages. Types of ANOVA Explained: The presenter differentiates between a one-way between-groups ANOVA and a within-subjects ANOVA, clarifying that the focus is on the former, which is simpler and more commonly used in research involving different groups. Understanding Data Requirements: For one-way ANOVA, the video explains the need for a continuous dependent variable (in this case, salary) and a categorical independent variable (job category). These are foundational for the analysis. Setting Up Data in SPSS: The tutorial includes detailed steps on how to access and set up the necessary data within SPSS. The presenter uses example data available within SPSS, which viewers can also access to follow along and practice. Configuring the ANOVA Test: The steps to perform the ANOVA test in SPSS are demonstrated. This includes selecting the dependent variable (salary) and the independent variable (job category), and configuring the analysis options like tests for homogeneity of variance and the plotting of means. Statistical Hypotheses: The hypothesis tested is whether there are significant differences in salaries across three job categories: clerical, custodial, and managerial. The expectation is that managerial staff would earn more than clerical and custodial. Exploring SPSS Menus and Options: The video provides a walkthrough of navigating the SPSS interface, focusing on how to enter variables into the analysis module and select appropriate statistical tests and options. Initial Results and Descriptive Statistics: Initial outputs from SPSS, including mean salaries and standard deviations for each job category, are shown. These help provide a preliminary understanding of the data before delving into deeper analysis. Discussion on Fixed and Random Effects: A brief discussion on fixed versus random effects in ANOVA is included, explaining that most analyses assume fixed effects. The presenter plans to elaborate on this topic in future videos. Plans for Further Analysis: The presenter mentions that post-hoc tests, which are crucial for identifying specific group differences after finding a significant F-statistic, will be covered in a subsequent video. Closing Remarks: The video ends with a reminder of the utility of ANOVA in statistical analysis and encourages viewers to experiment with the example data in SPSS to enhance their understanding of one-way ANOVA. This expanded summary provides a comprehensive overview of the tutorial, including both the procedural aspects of conducting an ANOVA in SPSS and the conceptual understanding necessary to interpret the results effectively.
I have a 1x2 chi square test wearing sunglasses which all subjects do vs eye color (light or dark) out of 200. I can’t get p value on calculator. 53 light eye wearing sunglases and 147 dark eye wearing sunglasses??
Great video! Thanks a lot! Wondering if it makes any sense to adjust for multiple comparisions when you are comparing means of a single variable between different groups e.g. pH values between 5 water treatments