Samenvatting
Summary (Research & Statistics - 7U9X0)
Samenvatting voor het vak "Research & Statistics" (7U9X0). Omvangrijke samenvatting van het hele vak.
[Meer zien]Voorbeeld 3 van de 18 pagina's
Voorbeeld 3 van de 18 pagina's
In winkelwagengesponsord bericht van onze partner
Verkoper
Volgen
Voorbeeld van de inhoud
Research & Statistics Summary Thom Bindels - 2021Research & Statistics Summary
Research Methods
Problem Formulation & Conceptual Model
Problem formulation: Research is always based on a certain problem, which can come from practice
or theory. The goal of the research is to provide an answer in the form of insight into the problem or
possible solutions. However, it is not the goal of research to provide the actual solution, only to supply
extra knowledge to come to the best possible solution. To do so, two steps need to be taken:
1. A research objective must be determined, this is basically why the research is necessary.
What is the purpose of the acquired knowledge from the research? A clear objective is
required so that the research can be set up properly.
I.e.: Objective is to gain insight in which factors influence a certain appreciation,
this knowledge is then used to find a solution that maximally increases a said
appreciation.
2. The objective is a certain statement, whereas also a research question is required. This
question asks what knowledge and/or insights must be generated. A research question is
accompanied by subquestions, which are the components of which the research question
consists of.
I.e.: Which factors influence an individual’s appreciation?
Question types: A question can be of a certain type, examples are frequency, difference or relation
questions. Quantitative research often uses frequency. Relation or difference questions can be more
relevant for qualitative research. Therefore, depending on the research type the question type
changes as well, this has an effect on how the data needs to be collected and analyzed.
1. Frequency questions allow for descriptive analyses and can be shown in graphs such as pie
charts, histograms, frequency tables, etc.
2. Difference questions have two variables, one is the groups to be compared and the other
variable is what is compared. This can be shown in boxplots and cross tables.
3. Relation questions have two variables in which the correlation between the variables is
relevant. A correlation can be positive in which a high score for one is a high score for another,
a negative correlation means that a higher score in one variable results in a lower variable for
the other.
Conceptual model: The model represents the variables (what) of the research and the corresponding
relationships (how) between these variables. These relationships are expectations and in statistical
analysis it is determined whether the relation is actually significant. The initial relationships are based
on experience, literature, preliminary research, etc. The model can also be used to determine the
subquestions the other way around, or an iterative process between the model and questions.
,Research & Statistics Summary Thom Bindels - 2021
All boxes in the conceptual model must be a variable. Not all links between the variables need to have
a direction, this can imply an association. This means that there is a relation, but it is not yet known
which variable is the cause and which is the effect. A conceptual model often has the main variables
of interest centrally, to which all the relationships point. All the other factors that influence this main
variable are distributed around them with these relationships point toward the main variable. The
conceptual model is related to the research questions. The dependent variables correspond to the
subjects of the main research question, and the relationships correspond to the various subquestions.
Check for each conceptual model the following:
1. Are the dependent variables the subject of the main research question?
2. All variables of the model occur as a subquestion (what/how much)?
3. All relationships occur as a subquestion (relationships between/influence on)?
4. The unit of analysis of all variables is consistent?
5. Causality is guaranteed for all relationships?
Unit of analysis: All the variables have an entity to which it relates, for the entire conceptual model
this must be the same. For example, this could be a property of an individual, this is often the case in
social sciences as this is oriented on persons.
I.e.: The gender of the individual, the frequency the individual uses the bus, the connectivity
that the individual has, etc.
Causality: Each relationship must be causal; this means that X (predictor) has an influence on Y
(explained) and X happens before Y. Often pitfalls of research are related to causality. This is often
related to spurious relationships, X and Y are both related to a common 3rd variable which removes
the correlation between X and Y. The conditions for causality are:
1. There must be a statistical correlation, which can be measured.
2. The cause must come before the effect, the time dependency can be difficult to determine.
3. Measure all possible causes to avoid spurious relationships.
4. Provide a theory that explains the causal mechanism, how does the cause generate an effect?
, Research & Statistics Summary Thom Bindels - 2021
Research Design
Research design: Choosing the best strategy to answer the research questions while minimizing
alternative explanations. Two important “ideal” types are:
1. The experiment, which allows for total control over the independent variables. Typically, one
group receives treatment (experimental group) and another group does not (control group).
The control group receives a “placebo”, which means the group believes it receives the
treatment but in reality, they do not. This is best used to rule out alternative explanations and
statistical analysis is rather simple. This is used to measure treatment effects and evaluative
research to test a hypothesis.
a. The Solomon design is a specific type of experiment that should be applied to any
experiment to make it successful. First, the participants of the experimental and
control group are randomly distributed, assuring an equal distribution of the sample.
Additionally, before and after measurements are performed for both groups to
exclude any autonomous changes that are not related to the treatment itself.
2. A survey typically provides a questionnaire to a large group of respondents. The research has
less control and therefore alternative explanations need to be ruled out by more advanced
statistical analyses. This is often used for explorative and descriptive research. It has the
following differences from an experiment:
a. All respondents receive the same questions and supply the same responses, no
control groups are possible.
b. Only one moment in time it is measured, this can give rise to problems on time
control. This lack of time-control make it impossible to identify causes and results,
therefore theory is required to determine the causality.
c. It supplies large sets of results (responses), this can be used for reliable measurements
(means). Moreover, many variables are researched so that they can be tested on
alternative causes (spurious relationships).
So why use a survey if it is less reliable than an experiment? Predictors sometimes cannot be
controlled, such as socio-demographics of persons. Additionally, there can be non-observable
variables (intentions, emotions, motives, etc.) that offer explanations for behavior or things
happened in the past.
Validity: This determines whether what has been measured is what is intended to be measured. There
are two types of validity:
a. Internal validity is about the internal consistency. Are the causal interpretations valid and are
alternative explanations ruled out? Experiments score high on internal validity, as everything
is controlled, and causality is no problem. In surveys it is difficult to exclude this causality.
b. External validity is about how generalizable the research is. If the research were to be
conducted at another time, place or within a different group, is it to show the same results?
For surveys external validity is often better, as they are more representable of the general
population. In an experiment an artificial environment is created by the researcher, which is
suitable for that specific research only.
Level of measurement: Measuring is assigning numbers to empirical phenomena. The level of
measurement determines how the numbers can be interpreted, and thus which analysis techniques
can be used as not all techniques can be used for all measurement levels. The first two levels are
discrete (non-metric) numbers providing qualitative information. The higher two levels have
continuous (metric) numbers providing quantitative information. The higher the level the more
analysis techniques may be applied; these techniques have larger power as well. The four levels are:
1. Nominal level is simply a distinction, there is no ordering between the numbers, and it is fully
random. A special type of nominal is dichotomous, where only two categories are such as
gender.
2. In the ordinal level, there is ordering present between the different variables. However, the
intervals between the levels are not known. An example can be education level.
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 ThomBindels. Stuvia faciliteert de betaling aan de verkoper.
Zit ik meteen vast aan een abonnement?
Nee, je koopt alleen deze samenvatting voor €4,99. Je zit daarna nergens aan vast.
Is Stuvia te vertrouwen?
4,6 sterren op Google & Trustpilot (+1000 reviews)
Afgelopen 30 dagen zijn er 56326 samenvattingen verkocht
Opgericht in 2010, al 14 jaar dé plek om samenvattingen te kopen