Quiz
307 test 2
This quiz has question types that open in the usual player.
Questions in this quiz (66)
1pre-experiments differ from true experiments as they do not include a true [blank_start]control group[blank_end]
- Acontrol group
2[blank_start]sample selection[blank_end] in quasi-experiments threaten a studys external validity because it is self selceted
- Asample selection
3threats to internal validity
[blank_start]testing effects[blank_end] - measuring the dependent variable at time 1 might casue change at time 2
[blank_start]maturation effects[blank_end] – changes that happen to the participant between measurements
[blank_start]history effects[blank_end] – events that occur between the measurements may account for the change
[blank_start]statistical regression towards the mean[blank_end] - extreme scores tend to fall closer to the mean when you measure them again
[blank_start]instrument decay[blank_end] – the measurement itself may change from pre to post test
[blank_start]sample selections[blank_end] - self selected and not random- Atesting effects
- Bmaturation effects
- Chistory effects
- Dstatistical regression towards the mean
- Einstrument decay
- Fsample selections
- Gmaturation effects
- Hsample selections
- Iinstrument decay
- Jstatistical regression towards the mean
- Khistory effects
- Ltesting effects
- Mhistory effects
- NTesting effects
- Omaturation effects
- Pstatistical regression towards the mean
- Qinstrument decay
- Rsample selections
- Sstatistical regression towards the mean
- Tsample selections
- Uinstrument decay
- Vhistory effects
- Wmaturation effects
- XTesting effects
- Yinstrument decay
- ZTesting effects
- AAmaturation effects
- ABhistory effects
- ACstatistical regression towards the mean
- ADsample selections
- AEsample selections
- AFmaturation effects
- AGstatistical regression towards the mean
- AHTesting effects
- AIinstrument decay
- AJhistory effects
4two types of pre-experimental designs are [blank_start]one shot case[blank_end] study and [blank_start]one group pre test post test[blank_end] study
- Aone shot case
- Bone group pre test post test
5[blank_start]one shot case[blank_end] study has the problem of there being nothing to compare the dependent variable scores to, so we cant tel if the DV was affected by the IV
- Aone shot case
- Bone group pre test post test
6an [blank_start]interrupted time series[blank_end] design is with only one sample, but takes measurements of the dependent variable multiple times.
- Ainterrupted time series
7three problems with a time series design are -
[blank_start]testing effects[blank_end] - measuring the dependent variable at time 1 might cause the change at time 2.
[blank_start]instrumentation effects[blank_end] - the measurement itself may change from pre to post test
[blank_start]participant mortality[blank_end] - participants drop off the study- Atesting effects
- Binstrumentation effects
- Cparticipant mortality
8what is less likely to happen in an interrupted time series design is -
[blank_start]history effects[blank_end] - events that occur between the measurements may account for the change
[blank_start]statistical regression to the mean[blank_end] - extreme scores tend to fall closer to the mean when you measure them again- Ahistory effects
- Bstatistical regression to the mean
9three problems of a non-equivalent control group design are -
[blank_start]statistical regression of the mean[blank_end] - extreme scores tend to fall closer to the mean when you measure them again
[blank_start]sample selection[blank_end] - pre selected rather than randomly selected
[blank_start]history effects[blank_end] - events that occur between the measurements may account for the change- Astatistical regression of the mean
- Bsample selection
- Chistory effects
10you are looking for a [blank_start]interaction[blank_end] between treatment condition and time of measurement in non-equivalent control group designs
- Ainteraction
11[blank_start]cross sectional[blank_end] methods examines multiple groups from different time cohorts whereas [blank_start]longitudinal[blank_end] follows one cohort over time
- Across sectional
- Blongitudinal
12being expensive and takes time to yield results is two disadvantages of the [blank_start]longitudinal[blank_end] method
- Alongitudinal
13being subject to cohort effects is the main disadvantage of the [blank_start]cross sectional[blank_end] method
- Across sectional
14common response formats for questionnaires -
[blank_start]categorical[blank_end] response - yes/no
[blank_start]rating[blank_end] scale - 1-5
[blank_start]ranking[blank_end] format - rank all options- Acategorical
- Brating
- Cranking
15common wording problems.
[blank_start]Vague/ambiguous terminology[blank_end] – items phrased vaguely can’t be sure what the responses mean. Ill- defined items are a similar problem.
[blank_start]Technical terminology[blank_end] – not common knowledge
[blank_start]Hypothetical questions[blank_end] – must appear reasonable to participants
[blank_start]Leading questions[blank_end] – may be suggestive
[blank_start]Value judgements[blank_end] – item wording should not contain implicit value judgements
[blank_start]Context effect[blank_end] – subtle effects on responses that depend on the nature of the rest of the questions
[blank_start]Double-barrelled questions[blank_end] – 2 questions in one
[blank_start]Hidden assumptions[blank_end] – assuming something in the question
[blank_start]Sensitive issues[blank_end] – questions may shock, embarrass, offend or annoy respondents
[blank_start]Social desirability[blank_end] – people tend to present themselves in a positive light- AVague/ambiguous terminology
- BTechnical terminology
- CHypothetical questions
- DLeading questions
- EValue judgements
- FContext effect
- GDouble-barrelled questions
- HHidden assumptions
- ISensitive issues
- JSocial desirability
16crowne-marlowe scale and the paulhaus deception scale are tests used to measure [blank_start]social desirability[blank_end]
- Asocial desirability
17three components of attitudes -
[blank_start]Affective[blank_end] – feelings about the attitude object
[blank_start]Behavioural[blank_end] – predisposition to act towards the attitude object in a certain way
[blank_start]Cognitive[blank_end] – beliefs about the attitude object- AAffective
- BBehavioural
- CCognitive
18the ideal properties of a scale designed to measure attitudes.
[blank_start]Unidimensional[blank_end] – measures only one trait at a time
[blank_start]Reliable[blank_end] – responses from the same person about the same thing should be the same
[blank_start]Valid[blank_end] – should measure what it claims to measure
[blank_start]Linear[blank_end] – people who score twice as highly should have twice as strong an attitude
[blank_start]Reproducible[blank_end] - respondents should make as few errors as possible- AUnidimensional
- BReliable
- CValid
- DLinear
- EReproducible
19a [blank_start]construct[blank_end] is a trait that is not directly observable
- Aconstruct
20questionnaire equation ---
[blank_start]observed score[blank_end] = [blank_start]true score[blank_end] + [blank_start]error[blank_end]- Aobserved score
- Btrue score
- Cerror
21a high [blank_start]cronbachs alpha[blank_end] level indicates the two items are measuring the same thing.
[blank_start]negative covariances[blank_end] will lower the score
a lot of [blank_start]items[blank_end] will increase the score- Acronbachs alpha
- Bnegative covariances
- Citems
22[blank_start]Dynamic[blank_end] traits change depending on context and experience
[blank_start]Static[blank_end] traits are relatively stable from time to time- ADynamic
- BStatic
23ways to measure the validity of a questionnaire/test ---
[blank_start]Content[blank_end] validity – experts asses the content of the questionnaire checking the readability, clarity and comprehensiveness.
[blank_start]Criterion-related[blank_end] validity – score on current scale is correlated with another measure. [blank_start]Concurrent[blank_end] – the score is compared with a criterion measure obtained at the same time. [blank_start]Predictive[blank_end] – score is correlated with a criterion assessed later. [blank_start]Postdictive[blank_end] – not very common, corelate scores you achieve now with something that has already happened.
[blank_start]Construct[blank_end] validity – degree to which an instrument measures the trait or theoretical construct- AContent
- BCriterion-related
- CConcurrent
- DPredictive
- EPostdictive
- FConstruct
24[blank_start]Construct[blank_end] validity is more concerned about how the scale fits with the theory about the construct whereas [blank_start]criterion-related[blank_end] validity is more concerned whether different measurements are related to each other.
- AConstruct
- Bcriterion-related
25we use the [blank_start]corrected item-total correlation[blank_end] as it excludes itself from the total
you remove an item from the [blank_start]cronbachs alpha if item deleted column[blank_end] if it is higher than the cronbachs alpha total
- Acorrected item-total correlation
- Bcronbachs alpha if item deleted column
26content analysis is considered to be a quantitative method as it uses objective, and systematic, counting and recording procedures. it involves [blank_start]random sampling[blank_end], [blank_start]measurement[blank_end] and [blank_start]operational definitions[blank_end] to turn content into numbers to be analyzed.
- Arandom sampling
- Bmeasurement
- Coperational definitions
27four characteristics of the content of text that may be used for coding -- [blank_start]frequency[blank_end], [blank_start]direction[blank_end], [blank_start]intensity[blank_end] and [blank_start]space[blank_end]
- Afrequency
- Bdirection
- Cintensity
- Dspace
28When measuring you should use [blank_start]structured observation[blank_end] and categories should be [blank_start]mutually exclusive[blank_end] and [blank_start]exhaustive[blank_end]
- Astructured observation
- Bmutually exclusive
- Cexhaustive
29[blank_start]manifest coding[blank_end] - when you code the visible, surface content in the text
- Amanifest coding
30[blank_start]latent coding[blank_end] - looks for underlying, implicit meaning in the text
- Alatent coding
31[blank_start]Manifest coding[blank_end] is very reliable because its either absent or present but multiple meanings of the word limits [blank_start]validity[blank_end]. [blank_start]Latent coding[blank_end] is less reliable as it depends on coders knowledge of context, language, and social meaning. Training improves reliability. [blank_start]Validity[blank_end] is better than manifest coding as it takes context into account.
- AManifest coding
- Bvalidity
- CLatent coding
- DValidity
32using [blank_start]stratified sampling[blank_end] ensures similar number of comments are directed at different demographic sub groups
- Astratified sampling
33three levels of analysis - Glass
[blank_start]primary[blank_end] analysis - the original analysis from a study
[blank_start]secondary[blank_end] analysis - re-analysis of original data from a study using different statistical techniques
[blank_start]meta[blank_end] analysis - statistical analysis of results from a large number of individual research studies so as to integrate their findings- Aprimary
- Bsecondary
- Cmeta
34[blank_start]meta analysis[blank_end] - combines empirical results from quantitative studies
- Ameta analysis
35[blank_start]power[blank_end] - the probability of detecting an effect when there really is an effect
- Apower
36[blank_start]effect size[blank_end] - the most commonly used test statistic used in meta analysis
- Aeffect size
37[blank_start]residual[blank_end] variance - variance due to sampling error calculated and subtracted from the observed variance
- Aresidual
38[blank_start]Homogeneous[blank_end] – if the residual variance is close to zero, we assume the effect is consistent across studies/situations and not moderated by features of those studies/situations
[blank_start]Heterogeneous[blank_end] – if the residual variance is high, the effect is assumed to vary in different situations.- AHomogeneous
- BHeterogeneous
39you use a [blank_start]fixed effect model[blank_end] in a situation where a low residual variance. it assumes that the effect we are studying is similar for all people in the population because the population is homogeneous.
- Afixed effect model
40you use a [blank_start]random effect model[blank_end] when residual variance is high assuming that the effect may vary for different people in the population and therefore heterogeneous.
- Arandom effect model
41[blank_start]fixed[blank_end] effect model - The effect sizes are weighted according to sample size such that studies with larger sample sizes are given more weight than small
- Afixed
42[blank_start]random[blank_end] effect model - Studies with different sample sizes are given a more equal weighting because variance is also used.
- Arandom
43[blank_start]random[blank_end] effects model is preferred over [blank_start]fixed[blank_end] effects model in most circumstances because we cant be certain that different studies are functionally equivalent
- Afixed
- Brandom
44[blank_start]sensitivity analysis[blank_end] - determines the robustness of the results in a meta analysis
- Asensitivity analysis
45weaknesses of a meta analysis -
[blank_start]publication bias[blank_end] - journals biased to what they publish and meta analysis only include published results
[blank_start]p hacking[blank_end] - researchers subtly manipulating their data so that effects reach significance
[blank_start]selection bias[blank_end] - selection of studies in order to produce a particular result- Apublication bias
- Bp hacking
- Cselection bias
46when you need to analyse nominal data as the dependent variable you use a [blank_start]chi square[blank_end] test
- Achi square
47chi squares compare whether an [blank_start]observed[blank_end] breakdown of scores across various categories fits some [blank_start]expected[blank_end] breakdown
- Aobserved
- Bexpected
48two types of chi square tests -
Chi square test for [blank_start]goodness of fit[blank_end] – use when there is only one nominal variable
Chi square test for [blank_start]independence[blank_end] – use when there are two nominal variables- Agoodness of fit
- Bindependence
49one-variable X2 test -
[blank_start]H0[blank_end] – the observed and expected frequencies do not differ. The sample comes from a population with the same distribution as the expected frequencies
[blank_start]Ha[blank_end] – the observed and expected frequencies differ. The sample comes from a population with a different distribution to the one the expected frequencies come from- AHo
- BHa
50how do we find the weighted mismatch
- Awe deduct the expected from the observed
- Bwe square the differences
- Cwe divide the squared difference by E
- Dwe deduct the expected from the observed
- Ewe square the differences
- Fwe divide the squared difference by E
- Gwe divide the squared difference by E
- Hwe square the differences
- Iwe deduct the expected from the observed
51calculations for df for a X2 test is Number of [blank_start]categories[blank_end] - 1
- Acategories
52the [blank_start]expected[blank_end] values are calculated in the ine variable X2 case as shown in the picture
- Aexpected
53chi square formula -
the square of the differences between [blank_start]expected[blank_end] and [blank_start]observed[blank_end] frequencies, divided by the [blank_start]expected[blank_end] frequencies and these are then summed for every cell
- Aexpected
- Bobserved
- Cexpected
54your X2 obtained is larger than the X2 critical you should [blank_start]retain[blank_end] the Ho
- Aretain
- Breject
55the null hypothesis for a X2 test of independence is [blank_start]Ha[blank_end] - the variables are related to each other
- AHa
5611. When a X2 test of independence is statistically significant, we conclude the variables are independent - [blank_start]Ho[blank_end]
- AHo
57the image is an example of working out the [blank_start]expected[blank_end] values for a X2 test of independence
- Aexpected
58[blank_start]phi coefficient[blank_end] - It is the measure of association for a 2x2 contingency table (smaller). Ranges from 0-1 and is like a correlation coefficient.
- Aphi coefficient
59[blank_start]cramers phi[blank_end] - The same an a phi coefficient but for larger contingencies
- Acramers phi
60one assumption that should be met in a X2 test is each score must not have any special relation to any other score so the observations must be [blank_start]independent[blank_end] of each other
- Aindependent
612x2 contingency tables expected cell frequencies should [blank_start]exceed[blank_end] 5 because If they are very small they have reduced [blank_start]power[blank_end].
- Apower
- Bexceed
62expected cell frequencies in [blank_start]larger[blank_end] contingency tables should be greater than 1 and no more than 20% should be less than 5.
- Alarger
63use a [blank_start]fishers exact[blank_end] test is expected cell frequencies are too small
- Afishers exact
64the more [blank_start]degrees of freedom[blank_end], the less the [blank_start]power[blank_end]. to maximise [blank_start]power[blank_end], use as many participants as possible with as simple contingency table as possible
- Adegrees of freedom
- Bpower
- Cpower
65three types of observation -
[blank_start]Casual[blank_end] observation – often at planning stage
[blank_start]Formal[blank_end] observation – unobtrusive observer making notes, visualy recording events and behaviour
[blank_start]Participant[blank_end] observation – observer is part of the events.- ACasual
- BFormal
- CParticipant
66four dimensions of observational research -
[blank_start]Theory testing[blank_end] --- [blank_start]exploratory[blank_end] – extent to which the study is testing a theory or exploring an issue
[blank_start]Experimental[blank_end] --- [blank_start]naturalistic[blank_end] – observation in the lab compared to in the ‘real world’
[blank_start]Structured[blank_end] --- [blank_start]unstructured[blank_end] – decide what to measure beforehand compared to no plan
[blank_start]Participant[blank_end] --- [blank_start]non participant[blank_end] – the complete [blank_start]participant[blank_end] – completely involved with group and conceals observations
[blank_start]Participant[blank_end] as an [blank_start]observer[blank_end] – have a meaningful social role in the group, observational activity not wholly concealed.
[blank_start]Observer[blank_end] as an [blank_start]participant[blank_end] – join the group with the intention to observe
Complete [blank_start]observer[blank_end] – detached from group of study, no direct contact- ATheory testing
- Bexploratory
- CExperimental
- Dnaturalistic
- EStructured
- Funstructured
- GParticipant
- Hnon participant
- Iobserver
- Jparticipant
- KParticipant
- Lobserver
- MObserver
- Nparticipant
Information
- Description
- psychological research Quiz on 307 test 2, created by Karissa Paterson on 18/09/2018.
- Author
- Karissa Paterson
- Language
- English
- Views
- 0
- Tags
- Rating
- No ratings yet Rate this quiz