Alle lectures & workshops van block 7 - Onderzoeksproject
10 views 0 purchase
Course
Onderzoeksproject (BK2103)
Institution
Erasmus Universiteit Rotterdam (EUR)
Alle workshops van Corrine en lectures van Romain worden in dit document besproken. Heel compleet document van alles wat is uitgelegd. Engelse deel zijn de lectures, Nederlandse deel de workshops in R. Heel uitgebreid beschreven in zo'n 60 pagina's, voor het laatste examen van dit vak (eerder midte...
- Lecture 1 = summary of block 6
- Lecture 1 & 2, Workshop 2 = Experimental design & research models with > 2 variables (e.g.
moderation)
o Why experiments are useful, the boundaries of correlation analysis, the need of multivariate
analysis
- Lectures 3-7 & Workshops 3-7 = Multivariate methods, specifically regression
o Background of regression analysis (3), application of regression analysis (4), advanced issues
with regression (5-7)
- Lecture 8 & Workshop 8 = presenting research results
This block = tutorial on how to do your research (thesis).
Summary of block 6
Definitions: from concept to variable.
Concept and constructs vary based on their degree of abstractness: a concept is MORE abstract than a
construct.
VB. Communication skill = CONCEPT. Vocabulary and syntax skill = CONSTRUCT. Is nog steeds niet meetbaar!
Nog steeds abstract. Beiden zijn abstract, niet direct meetbaar → oparationaliseren = VARIABLE. VB. 10
multiple choice vragen om vocabulary te testen (GMAT, mid-term exam, etc.)
Variables: from measurement to statistical analyses
1. Challenge 1: hoe vind je meetbare variabelen en meet je die?
2. Challenge 2: hoe interpreteer je de variabelen in het research model?
Level of measurement:
1. Nominal – met verschillende categorieën. VB. houd je van Duvel JA/NEE
2. Ordinal – hoe rank je de variabele op basis van je voorkeuren. VB. Biersmaken op volgorde zetten van
like
3. Interval – hoe rank je de variabele op basis van een schaal
4. Ratio – 0 heeft betekenis in de variabele. VB. hoeveel Duvels heb je gedronken deze week?
Special case of multi-dimensional scales
Hoe ziet de vraagstelling eruit? CASE 1
- One variable with several questions in a survey → beter known as ITEMS = questions
VB. meten brand loyalty met 4 vragen op likert scale “BR1, BR2, BR3, BR4”
Hoe combineer je deze 4 items over 1 variabele in je statistical analysis?? → mean score / average
bepalen
Hoe ziet de database eruit? CASE 1
- One item = 1 column
56
, - Check scale reliability with Cronbach alpha = de reliability voor meer dan 2 items!
- Alpha moet groter of gelijk aan 0.70 zijn.
- Dus bij a < 0.70 voor alle 4 de items, mag je items die groter of gelijk aan 0.70 zijn deleten
CASE 2 -> als er reversed-coded items zijn
VB. meten van brand loyalty met de vraag: I am NOT committed to this brand -> likert scale meet in the
opposite direction.
- You need to recode all the items:
o Old values = 1, 2, 3, 4, 5, 6, 7
o New values = 7, 6, 5, 4, 3, 2, 1
Zie R-sheets onder Block 7 → R1.
In R-sheet maak je een nieuw dataframe waar alle waarden boven alhpa 0.70 zijn. Alpha = 0.84. Er zijn 4
kolommen in je nieuwe dataframes.
Welk item zou je verwijderen als je de HOOGST mogelijke reliability (alpha) zou willen hebben → Item nummer
BR4!
Regel 31 in R vormt de likert scale om met nieuwe values.
Hoe bereken je de mean van 4 categorieën? -> nieuwe variabele creëren in regel 38.
Dus:
1. Eerst Alpha Cronbach berekenen op alle items (5) -> welke komt onder de alpha 0.70? Dan moet je
items verwijderen
2. Als je een negatief gestelde vraag hebt (ik ben NIET loyaal) -> dan moet je deze eerst recoden
3. Dan het gemiddelde van die 5 items berekenen
4. Hypothese verwerpen of niet
Waarom zou je een negatief gestelde vraag in je survey zetten? Om te checken of mensen paying attention!
Anders kan je telkens 6 aanvinken en val je bij een negatieve vraag door de mand.
CASE 3 -> items zijn niet gemeten op dezelfde schaal
- Rescale the items!
- VB. schale 1-5 and scale 1-7
- Compute a new scale (zie slides!)
- Cronbach alpha
CASE 4 -> one interval and one nominal -> more categorical questions to one variable
DEZE KAN JE NIET MERGEN!
Het is alleen mogelijk als: aggregating nominal questions
- Alle items binary zijn, correct of niet correct antwoord
- Als de antwoorden consistent zijn, bijv. altijd TRUE/ FALSE, bijv. brand loyalty is altijd TRUE
- Aggregated measure at individual level, how many is “right”? → RATIO
Nominal = het antwoord is ofwel TRUE ofwel FALSE
57
,Challenge 2 -> several variables within the same research model
- Theoretically: build an empirical model to research the relationships between the variables
- Empirical: …
Simplest research model has 2 variables
- IV – wearing glasses
- DV – grade
Theoretical hypothesis about the relationship between the variables -> moet wel een direction bevatten in de
hypothese, bijv. Students who wear glasses have higher grades.
Researchers often label their theoretical hypotheses as following:
- H1: X positively influences Y
- H2: X negativily influence Y
For each of these theoretical hypothesis we can associate two statistical hypothesis known as null vs.
alternative:
- H0 (statistical)
- H1 (theoretical)
Confusion with terminology and notations
- In almost all research reports, researches DO NOT report H0 or H1 → only mention if it’s confirmed/
rejected/ etc.
- What you should state “Results from statistical analyses show that the correlation coefficient between
X and Y was positive (r=0.21) and significant (p=0.01). H1 is validated”
Dus H0 en H1 moet je weten om onderzoek te doen, maar hoef je niet in je rapport te zetten.
VB. H1 = variable X positively influences variable Y → NIET H1 = p < 0.05 bijvoorbeeld.
Choose right statistical test
1. T-test
2. Correlation
3. Chi-squared
4. ANOVA
(2) Correlation
Correlation coefficient is known as row = p (Greek letter).
VB. H1: ease of use positively influences technology adoption
- H0: r < 0
- H1: r > 1
- In you research report you don’t have to write H0 and H1 this way!
- Results: r =0.21, p = 0.01
- P < alpha 0.05 dus reject H0
- State: “the results confirm the researcher’s hypothesis”
58
, (1) T-Test
Comparing average means between two categories.
VB. Average grades for people with or without glasses
- H0 = grades wearing glasses < grades no glasses
- H1 = grades wearing glasses > grades no glasses
Hypothesis is if wearing glasses will improve your grades (H1).
- Results: grades with glasses = 8.5 and grades without glasses = 7.4
- T(500) = 4.90, p < 0.001
- Look at the mean! The grades 8.5 > 7.4 → is it significant enough?
- P-value < 0.05
- Results confirm the research hypothesis
Stel de gemiddelde cijfers zijn andersom dus zonder glasses 8.5 → results do not confirm the research
hypothesis. Results actually show the opposite effect as hypothesized.
T(500) = degrees of freedom. Participants n -1.
(4) ANOVA
Compares average means between different groups!
VB in exam van
- IV: 3 categories: calorie label, no label, tri-color label
- DV: fruit consumption
- H1: the researcher hypothesize that only the tri-color label will increase fruit consumption
- Significance we CANNOT observe from the boxplot!
First step of ANOVA: General omnibus test/ F-test → do at least two groups have different average means
Dit is een table met alle gegevens van means van groepen
Stap 2 is Multiple comparisons → elke groep wordt vergeleken met de andere groep → alle mogelijke
combinaties worden gemaakt, 2 bij 2.
Zie laatste colom van p adj -> vergelijken met alpha om H0 te verwerpen/ aan te nemen.
VOORBEREIDING COLLEGE 18 -1-2021
1. In the first "online experiment" (as detailed in text and with the first figure with three bars). What is the
independent variable and the number of experimental conditions? What is the dependent variable?
Independent variable: De manier van de vraag stellen voor toestemming met neutrale/ opt-in / opt-out optie ->
dat mensen automatisch als donor worden ingeschreven of niet.
Number of experimental conditions: opt-in, opt-out, neutral (opt-in default = not an organ donor, confirm your
choice? / opt-out = you are an organ donor, do you confirm your choice? / neutral = do you want to be an
organ donor?)
Dependent variable: donateur zijn, donatiepercentages. -> do participants register as organ donor: JA/ NEE
59
The benefits of buying summaries with Stuvia:
Guaranteed quality through customer reviews
Stuvia customers have reviewed more than 700,000 summaries. This how you know that you are buying the best documents.
Quick and easy check-out
You can quickly pay through credit card or Stuvia-credit for the summaries. There is no membership needed.
Focus on what matters
Your fellow students write the study notes themselves, which is why the documents are always reliable and up-to-date. This ensures you quickly get to the core!
Frequently asked questions
What do I get when I buy this document?
You get a PDF, available immediately after your purchase. The purchased document is accessible anytime, anywhere and indefinitely through your profile.
Satisfaction guarantee: how does it work?
Our satisfaction guarantee ensures that you always find a study document that suits you well. You fill out a form, and our customer service team takes care of the rest.
Who am I buying these notes from?
Stuvia is a marketplace, so you are not buying this document from us, but from seller riannestals. Stuvia facilitates payment to the seller.
Will I be stuck with a subscription?
No, you only buy these notes for $5.83. You're not tied to anything after your purchase.
