Hello all,
I am doing a research project for class. Our research questions are as follows:
Q11)Has news exposure impacted parental decision-making in preparation for the new school year?
Q2) Has Media exposure to violence impacted parental decision-making in preparation for the new school year?
Q3) Have school shooting impacted parental decision-making in preparation for the new school year?
Three IVs we believe may predict DV. However, I do not understand the nuances of differences between t tests, f tests, exact tests, etc. Best guess I have is it is multiple regression? But what specifically? I am trying to determine sample size and will eventually also need to know how to analyze the data. Also, I believe the variables are categorical, based on self-report scales.
There are some issues here: first of all, It is not clear what is meant by "parental decision-making" - this is the response variable and determines the statistical part of the model. Further, it is not clear if the samples are random samples and what population is sampled, and if there is some co.variance structure in the sample (e.g. values from the same class/school/city/country/year) and if there are other omitted variables that are likely to have some impact (age of the pupils, parent's SOS, parent's education/profession, type of school, etc). And finally, a problem for interpretation will be that you will have observational data,not experimental data. This hinders the interpretation effects as bein causal. You can show or demonstrate correlations, but the data itself does not give you any evidence for causality.
Eventually you are correct: some multiple regression model will be the method of choice, but it remains the challange to find out which predictors to consider and if there are non-linearities and interactions involved, and which distribution of the response variable would be reasonable. After this is all settled, you can fit the model and test specific statistical hypotheses about relevant parameters (orcoefficients) in the model. Depending on the fact whether the model requires to estimate the variance as an independent parameter the tests of infividual parameters/coefficients will be either t- or z tests. It is also possible to compare nested models with either F- of Chi²- tests.
Hello Jochen,
So parental decision-making or more-so, school-related parental decision making is measured also through self-report. It's five questions that include likert scale answers. Questions like: how often have you considered changing your child's school or learning environment? how many items have you purchased that will help your child stay safe? Things like that. The population is non-random I believe. It is through a school project so we do not have a lot of resources and definitely even less time. I am conducting it through posting on Reddit, emailing it to two colleges with their permission, and through personal social contacts of researchers. Our moderating variables or descriptive statistics include: age, gender, political party, support of gun ownership. We're not going for causality with this current study.
Choosing the right test to compare measurements is a bit tricky, as you must choose between two families of tests: parametric and non-parametric. Many -statistical test are based upon the assumption that the data are sampled from a Gaussian (normal) distribution. These tests are referred to as parametric tests.
To select the statistical test, you should take into account some criteria such as number of variables, types of data/level of measurement (continuous, binary, categorical) and the type of study design (paired or unpaired).
- Badges
- Science topic
- Similar topics
- Thesis Research