week 5 - 7 - qualitative research

LO5.1 When and why qualitative research is conducted and identify different qual research approaches and their appropriate uses

• used to understand conceptual beliefs

• is a more naturalistic method

• used to create an intervention or understand the subjectivity of the properties

• understands meaning and cause

• is constructivist - knowledge based on experiences (positivist - observations (quantitative))

LO5.2 Understand the quality of qualitative research

• soft - not mathematical

• less understood

• less rigorous

• more confirmation bias

• more social desirability bias

• based on the skills of the interviewer

Critical Appraisal Program: 10 questions based on; Relevance, Rigour and Credibility

LO5.3 Explain how research questions and conceptual framework guides qualitative research

• Organises key variables, concepts and relations

• allows for development and inclusion of other variables

LO5.4 Key components of qualitative research design — sampling strategies, data collection methods and interview schedules

sampling

• convenience

• criterion - meets a criteria

• discriminatory or confirmatory — deeper understanding

• extreme case — unseen variables

• homogeneous

• max variation — ranges of behaviours and their connections

data collection

• structured interviews - standardised questions w/ same person

• semi-structured - key q - flexible

• unstructured - q areas and open ended

• in depth

• focus group - guided open discussion

• observational

must be interactive and objective

the number of participants needed determined by quality of info → informational power

interview schedule

• Prep - proffesional, on time, know your shit, know them

• Begin - what? why? how long? confidentiality, consent, understanding

• Rapport - comfort, tailored to them

• Warm up - choose Q wisely, get them talking, context, listen + be interested

• Questions - related, open, broad then specific, simple, don’t assume, probe

• End - thank and debrief

LO5.5 what is research reflexivity?

neutrality in research

avoiding bias

self- conscious critiquing and evaluation → explicitly about context, beliefs and sources of bias

LO6 overview of z tests and t tests

z - test - a comparison of the single sample mean to a value, when pop variance/ sd is known

t - test - a comparison of the mean of a signle sample with pop variance isn’t known

LO6 - what are biased and unbiased estimators?

unbiased - the sample means are unbiased estimators of the population mean > the mean of sample mean (sum of X(bar)/n) is equal to the true pop mean

biased - the sample vairance are biased estimators of the pop variance > the mean of sample variance : (sum(x-meanofx)²/n > this is only suitable for descriptive stats) is less than the pop variance

iwe need to make in unbiased by increasing SD

sum(x-meanofx)²/n-1 <is a degree of freedom

LO6 what is the degrees of freedom

number of observations = number of things being estimated

LO6 - what is the critical t and critical p approach

t - statistical - significance interval of 0.05 - usually directional, if t-value is more extreme then pop-t then we reject the null

p- value - 0.05, directional, if p > 0.05 then we reject

if non-directional - p=0.025

LO6 what is the point estimate and interval estimate

point - mean of x

interval estimate - mean of x ± error

LO6 what are confidence intervals

the percentage that confidently holds the true mean - usually 95%

width → increased sample size decreases the interval and makes it more precisise, decreased variation lowers the interval

+go hand in hand with the boundary of 0.05:

• if H0 is retained at 0.05, CiI contains the null

  • if H0 is rejected at 0.05, CI doesn’t contain the null

LO^ - what are the advantages and disadvantages of repeated measures…?

adv - a decrease in error, participants own their own control

Dis - carry over effects, not always feasible

what is counterbalancing?

systematic variation of the order of treatment to avoid confounds between order and treatment

• equal no. of times completed

LO6 - what are the errors in inference?

the truth exists that we don’t know, we have to decide

When the H0 is false and:

• we reject the H0 - correct (probability = 1-beta)

• we retain the H0 - Type II error - false negative (probability = beta)

When the H0 is correct and:

• we reject the H0 - Type I error - false pos (probability = alpha)

• we retain the H0 - Correct (probability = 1 - alpha)

Increase sample, decrease variance = increase precision for point estimate

One - tailed test - larger alpha

LO6 - what is effect size?

how well the experiment/intervention worked - higher means it is better or more effective

LO7 - One-way analysis of variance (considered revising sample means. assumptions and their violations)

Sample Means;

• Single sample in comparison - z-test (popsd known), t test (popsd unknown)

• Single sample measured twice - related samples t test

• two samples: independent samples t test

Assumptions:

• two groups are independent and randomly sampled >non - use a different analysis

• the dependent variable is measured on an interval scale > use a different analusis

• dependent variable is normally distributed > if equal n in groups - not a problem

• homogeneity of variance - equal variances in their sample (test: levene’s test) > only problem if n not equal

• interval level measurement

Violation of assumptions ^ > t test is robust in violations

LO7 - outline internal validty and it’s threats

The extent to which a study accurately establishes a causal relation between independent and dependent variables

• effect, cause, impact etc. - always investigate this language

threats:

• confounding variables

• maturation - participant changes over time

• testing effects - repeated testing may imporve performance

• regression to the mean - extreme values naturally move towards the average

random allocation - best resolution

LO7 - outline external validity

the ability of a study’s findings to apply beyond the specific conditions of the research

• generalisability - sample to pop

LO7 - what are the types of control?

• waitlist control - experience nothing different but a slight test through an questionnaire (should do slightly better than the normal control)

• active control condition - questionnaire and partial intervention

however this can inflate type 1 errors

LO7 - what is the analysis of variance → ANOVA

attempt to solve problem of dealing with multiple groups CONTINUE NEED TO UNDERSTAND AND I CAN"‘t RN