Samenvatting
Summary ARM: lectures, workgroups and literature
Summary ARM: lectures, workgroups and literature
[Meer zien]Voorbeeld 4 van de 64 pagina's
Voorbeeld 4 van de 64 pagina's
In winkelwagengesponsord bericht van onze partner
Verkoper
VolgenVoorbeeld van de inhoud
Summary Advanced Research MethodsWeek 1
Introductive e-learning
Quantitative research
Types of measurement levels of variables:
1. Nominal: The scale points are distinguishable (te onderscheiden), but there is no ranking.
E.g. place of birth or marital status. They can also be seen as outcome variables. Variables
with two scale points (yes/no) are called binary variables or dummy variables (the name not
the measurement).
2. Ordinal: the scale points are distinguishable and can be ranked but the distance between the
scale points is not necessarily the same. E.g. self-assessed health status (bad, average, good,
very good, excellent) or education level (low, middle, high). The difference between ‘bad’
and ‘average’ is not necessarily as big as the difference between ‘very good’ and ‘excellent’.
3. Interval: the scale points are distinguishable, can be ranked, the distance between the scale
points have a meaning but there is no absolute zero. E.g. temperature (Celsius/Fahrenheit)
or IQ.
4. Ratio: the scale points are distinguishable, can be ranked, the distance between the scale
points have a meaning and there is an absolute zero point. E.g. age, height, speed.
Outcome vs. exposure variables
Outcome variable: has the focus of attention, whose variation or occurrence we are seeking
to understand. Another name: dependent variable
Exposure variables (or identifying factors/determinants): may influence the size of the
occurrence of the outcome variable. Another name: independent variable
Examples of descriptive statistics: average/mean, median, frequency, percentage etc.
Number of issues which can lead to intentionally or accidentally misinterpreting data:
1. Precision vs accuracy
2. Unclear definitions
3. Unit of analysis
4. Mean vs median
5. Apples and oranges
6. Absolute and relative differences
7. Starting points and endpoints
8. Ignoring context
Qualitative research
Deductive coding: approaches usually begin with a theory-driven hypothesis, which guide
data collection and analysis. It is like creating analytical depth and connection to existing
literature. Theory 🡪 hypothesis 🡪 observation 🡪 confirmation
Inductive coding: begin with a research question and the collection of empirical data, which
are used to generate hypotheses and theories. It provides new insights, close to the actual
reality of respondents. Observation 🡪 pattern 🡪 tentative hypothesis 🡪 theory
Validity:
- Member check: present your analysis to the people you have studied so they can validate
your interpretation
- In depth- data collection: collect enough data to ensure you know what you are talking about
- Exploring deviant cases: look for disconfirming evidence and present it
, - Data triangulation: use multiple data sources: e.g. interviews with different groups combined
with observations
- Including context of the research for the reader to judge your analysis: include information
on the research project, topic list etc.
Reliability:
- Self-reflection of the researcher
- Accurate field notes and verbatim transcriptions of your interviews
- Discussing coding amongst researchers
- Including raw data in the presentation of your results
- Explaining how your data links to your interpretation
Generalizability:
- Use of sensitizing concepts
- Show how findings can be transferred to other settings
- Reflect on the limitations of your study
! The use of large representative sample is an example of generalizability of a quantitative
research not a qualitative research !
Ethical considerations:
- Consent: respondents of the interviews/observations need to know that they are being used
for your research. Before you record them, you must get their consent.
- If respondents want to withdraw from your study, they must have the opportunity to do so
at all times.
- Ensuring the privacy of the respondents. It is important to ensure anonymity of your
respondents both when saving as well as presenting your data. It is important to talk about
this with your respondents before your interviews/observations.
- Provide room for your respondents to tell their story and raise issues they think are
important instead of sticking to your predetermined ideas.
1.1 Lecture DAGs
DAGs help to visualize the causal structure underlying a research question.
RQ: What is effect of X on Y?
● Directed
o Connection between X and Y follows the direction of the arrows (X precedes Y in
time)
o An arrow means a possible causal effect
o No arrow means certainly no causal effect (is a heavier claim than drawing an arrow
that should not be there)
● Acyclic (‘cyclic’ = ‘goes round’, ‘acyclic’ = ‘cannot go round’)
o A path of arrows never comes to its origin (a variable cannot cause itself)
● Graph
Causality is used everywhere: what is the effect of, what is the impact of, what is the
influence of
,When you adjust for a confounder (so closing a backdoor path) it looks like this, with a box
around it.
Paths
● A
path is any
route between exposure X and outcome Y
● Paths do not have to follow the direction of the arrows (so every line from X to Y is a
path, no matter what the arrow looks like)
Causal and backdoor paths
● A causal path follows the direction of the arrows
● A backdoor path does not
Open and closed paths
● All paths are open, unless they collide somewhere on a path
● A path is closed if arrows collide in one variable on that path
Blocking open paths
● Open (causal or backdoor) paths transmit association
● The association between X and Y consists of the combination of all open paths between
them
● Here: all paths except X 🡪 W 🡪 Y
● An open path is blocked when we adjust for a variable (L) along the path
● This means that we remove the disruptive influence of L from the association between
X and Y or adjusting for a variable on that path in a regression analysis
● How? By including variable L in the regression analysis
● Backdoor paths always need to be closed
● Causal paths need to be open/closed depending on RQ
Opening blocked paths
● Including a collider (W) in the analysis means you open the blocked backdoor path (this with
disrupt the association between exposure X and outcome Y) and will cause bias
● To obtain an unbiased estimate you need to avoid this, by closing the backdoor paths again.
F.e. by removing the collidor from your analysis or with more variables on that backdoor
path).
Assessing causal inference studies (IMPORTANT!)
● What was the explicit research question?
o What was the implicit question?
● What was actually estimated?
o Is the estimate biased or unbiased?
o Is this an estimate of a full or partial effect?
● Is the estimate really an answer to the research question?
● How was the analysis designed?
● Which statistical methods were applied?
● Were these methods applied correctly?
● What is the (type of) estimate? Is it big, small, good, bad, etc.?
, o How uncertain is the estimate?
● What do the researchers conclude? Is that conclusion justified?
● Is the conclusion supported with strong or weak evidence?
● How does the conclusion compare to what we (already/thought we) knew?
Case with the L’Oréal powder:
Arguments against buying the powder:
● Small sample size (n=41) only tested on 41 women.
● You do not know how the women applied the powder, like how much they used it and how
often.
● Claim that it is the nr 1 powder but only 41 included, in which country? Can you generalize
this group?
● Study performed of financed by L’Oréal itself (commercial company)
● They do not specify the quality of the skin (oily, there are different).
● They did not use a control group
o Essential data is missing
o What would happen without treatment?
In causal inference:
● We are not interested in the outcome per se (i.e., 70% less imperfections), but …
● We are interested in the role of the treatment in achieving this outcome (i.e., without True
Match Minerals powder, would there have been less skin imperfections?)
🡪 We do not have this information so no causal claim can be made based on this study.
Formal definition by Hernàn and Robins (2020):
‘In an individual, a treatment has a causal effect if the outcome under treatment 1 would be
different from the outcome under treatment 2.’
Formal notation of causal effect: 𝑌𝑖𝑎=1 ≠ 𝑌𝑖𝑎=0
🡪 We would like to know, on an individual level, what would be the outcome if person Y
would have received the treatment, in comparison to the outcome for person Y if he/she has
not received the treatment. In order to estimate the effect of the treatment on an individual
level.
Y = outcome a = treatment
1 = yes (received treatment) 0 = no (received no treatment)
i = individual ≠ does not equal
● 𝑌𝐾𝑎=1 = 1 (improvement with treatment)
● 𝑌𝐾𝑎=0 = 0 (no improvement without treatment)
● Treatment effect for K: ΔYK = 1 – 0 = 1 (positive effect)
● Average treatment effect = average of ΔYi
A=1 (used the powder) / A=0 (not used the powder)
This is the ideal example of observation, but in rl you never know what you did not observe
or what would have happened if a specific person did not use the powder. We can however
try to see what would have otherwise happened using a control group with people of similar
backgrounds, age (exchangeable and the only difference between the groups should be one
group gets the treatment while the other doesn’t).
● Potential outcomes approach: comparing different potential outcomes of which we can only
observe one.
Voordelen van het kopen van samenvattingen bij Stuvia op een rij:
Verzekerd van kwaliteit door reviews
Stuvia-klanten hebben meer dan 700.000 samenvattingen beoordeeld. Zo weet je zeker dat je de beste documenten koopt!
Snel en makkelijk kopen
Je betaalt supersnel en eenmalig met iDeal, creditcard of Stuvia-tegoed voor de samenvatting. Zonder lidmaatschap.
Focus op de essentie
Samenvattingen worden geschreven voor en door anderen. Daarom zijn de samenvattingen altijd betrouwbaar en actueel. Zo kom je snel tot de kern!
Veelgestelde vragen
Wat krijg ik als ik dit document koop?
Je krijgt een PDF, die direct beschikbaar is na je aankoop. Het gekochte document is altijd, overal en oneindig toegankelijk via je profiel.
Tevredenheidsgarantie: hoe werkt dat?
Onze tevredenheidsgarantie zorgt ervoor dat je altijd een studiedocument vindt dat goed bij je past. Je vult een formulier in en onze klantenservice regelt de rest.
Van wie koop ik deze samenvatting?
Stuvia is een marktplaats, je koop dit document dus niet van ons, maar van verkoper fabinnegeerling. Stuvia faciliteert de betaling aan de verkoper.
Zit ik meteen vast aan een abonnement?
Nee, je koopt alleen deze samenvatting voor €9,48. Je zit daarna nergens aan vast.
Is Stuvia te vertrouwen?
4,6 sterren op Google & Trustpilot (+1000 reviews)
Afgelopen 30 dagen zijn er 66579 samenvattingen verkocht
Opgericht in 2010, al 14 jaar dé plek om samenvattingen te kopen