This includes the microlectures (B2-2) Choosing the right databases, (B2-3) Search strategy, and (B2-4) Documenting and working systematically

These are content for the exam of the Research Methods component of Module 7 (Research Methods and Research Project) of the University of Twente, in the Netherlands. 

Choosing database

• Search engines: list of results from the open web --> e.g. google
• Academic databases: search in limited body of scientific literature
  
  • It is advisable to use 2-4 databases
    
    • Relevant to your research
    • Matching information you need
    • Adding something unique

What is available:

• Search engines:
  • Little control over your search, too many results
• Specialized database: specialized in specific type of information --> e.g. Orbis, Espacenet, SciFinder
• Publisher database: databases from specific publishers --> e.g. ScienceDirect, SpringerLink
• Domain-specific database: search on specific discipline --> e.g. Business Source Elite, Philosopher’s index
• Multidisciplinary database: e.g. Scopus (includes more health sciences), web of science (older sources, high focus on life sciences and humanity)

Search strategy

Boolean operators

• AND: to look for results include all the terms combined by and
• OR: to look for anyone of the terms. Broader search than and
• NOT: to filter out

Boolean modifiers

• (Parentheses): to make sure search is in the right sequence
• “Quotations”: used for multiple worded terms
• Asterisk*: includes all words with the same root

Searching within your results:

• 1. Sorting: sort the order of the results (e.g. relevance)
• 2. Refining options: e.g. exclude results older than 2020 --> only refine because of specific reason
• 3. Snowballing: look into reference list (or related documents) of an article, and expend your research from there
  
  • MeSH:
    
    • 1. Type key word (e.g. stress)
    • 2. You’ll be provided with all types of stress
    • 3.  Choose your interest topic --> find list of synonyms that can be copied to search log
      
      • OR look up your interest topic adding [mesh]

Search matrix

• SPICE: setting, perspective, intervention, comparison, evaluation
• CIMO: context, intervention, mechanisms, outcome
• PICO: population, intervention, comparison, outcome
• PICo: population, interest, context --> e.g. healthcare workers, stress, COVID

This includes summaries of the publications preperation for Practical Assignment small story presentation of selves

These are content for the exam of the Research Methods component of Module 7 (Research Methods and Research Project) of the University of Twente, in the Netherlands. 

Article 1: ‘Define, Explain, Justify, Apply’ (DEJA): An analytic tool for guiding qualitative research sample size

• This article contributes a strategic and holistic approach to sample size determination in qualitative research via an analytical tool termed DEJA
  • DEJA: 4 step approach, guides users to articulate and understand the various considerations that influence sampling based on the unique characteristics of their study and research approach --> approaches the multiplicity of considerations in qualitative research sample size determination
    • Involves explicitly defining the factors influencing their sample sizes and justifying the decisions.
    • The DEJA overcomes the tendency towards quantifying qualitative research
• Sampling for qualitative studies:
  
  • 'Appropriate' sample size based on the concept of saturation (refers to reaching a point where data collection is considered satisfactory)
    
    • “Saturation is the point where no new categories, themes, or explanations emerge from the data”
  • Identifying factors that may influence sample size --> include researcher's knowledge, experience, subjective judgment, quality of information, sampling technique, research strategy, analysis technique, aims of the study, and more
    
    • The DEJA aims to make explicit the factors influencing sample size decisions and encouraging a holistic approach to determining sample sizes
  • A fixed predetermined sample size is seen as less than ideal in qualitative designs

DEJA: Holistic integrated concept for sampling in qualitative research

DEJA does not aim to provide a definitive or perfect sample size but rather guides researchers in the critical thinking process of selecting an appropriate sample for their specific study

• 1. Define: defining the sampling strategy --> involves drawing upon literature to articulate the chosen approach (e.g. non-probability sampling)
• 2. Explain: explaining the technique associated with the strategy (e.g. purposive or snowball sampling)
• 3. Justify: justifying the sampling decisions requires a comprehensive approach, aligning these decisions with the research's components (e.g. RQ, theoretical framework)
• 4. Apply

Conclusion

• This study does not aim to establish empirical guidelines for determining the "appropriate" sample --> it emphasizes the importance of embracing various techniques while keeping in mind the core principle of qualitative designs: understanding complex human issues
• The focus should be on gaining an in-depth understanding of phenomena through participants' lived experiences, rather than solely seeking generalizability or saturation of results
• DEJA reflects an awareness of the multidimensional nature of qualitative studies.

Article 2: Inductive, Analogical, and Communicative Generalization

• Generalizability: this article introduces criteria to improve the quality of generalization based on analogy
• 3 types of inductive generalization (process

This are the summaries of Chapter 10 and Chapter 11 of the Analysing Data using Linear Models Book (Van Den Berg, 2021)

These are content for the exam of the Research Methods component of Module 7 (Research Methods and Research Project) of the University of Twente, in the Netherlands. 

Chapter 10: contrasts

• ANOVA: make comparisons between means of different groups
  • In R --> when we run linear models, this is the default comparison:
    • Suppose compare the mean height in countries A, B and C --> run a linear model, and the first group (country A) becomes the reference group --> output:
      • Intercept = the mean of the reference group
      • Slope parameters: differences between countries B and A, and the difference between countries C and A
  • --> This chapter focuses on how to make choices regarding what comparisons you would like to make

Contrast:

• A contrast is a linear combination of parameters or statistic
  • Linear combination is a weighted sum, a regression equation like b1X1+b2X2 (a linear combination of independent variables X1 and X2, with weights b1 and b2)
    • Regression equation: a statistical model, used to determine the specific relationship between the predictor variable and the outcome variable
• Example: focus on the mean blood pressure of Dutch population (MDutch) and the mean blood pressure of German population (MGerman)

• • “A contrast is a weighted sum of group means”
    • In L1, the weights b1 and b2 equal 1 and 1 respectively
    • In L2, the weights b1 and b2 equal 1 and - 1 respectively
  • Row vector: L1 can also be displayed as  , and L2 as (the order changed)
  • Contrast matrix L: combine these two row vectors

Recap of previous chapters:

• Dummy variable: when one variable has multiple categories, these can be coded numerically and convey the same information
  • Example on blood pressure (bp_diastolic) with categorical variable nationality

• • • The slope represents the difference in the means for the two groups (german/dutch)

• • • Intercept is the average blood pressure in the Dutch group
    • The 1

These are summaries of the Chapter "Data Cleaning and Text Preprocessing" and Chapter "Text Clustering and Topic Modelling" (Maryam Amir Haeri and Stéphanie M. van den Berg)

These are content for the exam of the Research Methods component of Module 7 (Research Methods and Research Project) of the University of Twente, in the Netherlands. 

Data Cleaning and Text Preprocessing

Introduction to Big Data Analytics

• Big data analytics is a growing field due to social media platforms like Twitter and YouTube --> the diverse sources of big data contribute to its wealth of information and potential applications across various domains
  • Big data analytics is the use of advanced analytic techniques for very large and diverse data sets
• Big data refers to data in large quantities, complex, and generated fast so that it is hard or (even impossible) to process by utilising traditional approaches
  
  • Sometimes big data is characterised by 5Vs:
    
    • 1. Volume: big data is data containing lots of instances and variables
    • 2. Velocity: he speed of generating big data is often very high, and it is essential to have fast techniques for analysing them --> Otherwise, if the data processing takes too much time, the results may be outdated and no longer useful by the time the results are there
    • 3. Variety: we encounter various types of data that often need to be combined in one analysis --> moreover, data may be generated from various sources, where some of the data may be highly structured and some data may not
    • 4. Veracity: Most of the sources of big data are not fully trustworthy.
    • 5. Value: Processing any large dataset does not necessarily lead to valuable knowledge. But if the proper data analytic approach is used, a lot of money can be made --> e.g. cookies that collect data on online behaviour of many individuals can generate billions of Euros from advertising alone

The Big Data Analytical Process

• 1. Problem identification: what needs to be predicted, its significance, the level of accuracy required, and whether real-time results are necessary.
• 2. Selection of Data Resources: considering all potential data sources available and choosing the ones essential to solving the identified problem
• 3. Data cleaning: addressing issues such as duplicates, corruption, irrelevance, noise, and missing data by fixing or eliminating these instances.
• 4. Data transformation: converting the collected data into a suitable format for machine learning algorithms
• 5. Mining: selecting appropriate data analytic methods and configuring their parameters
• 6. Analysis of Results: evaluating and interpreting the outcomes of the data mining process using various criteria and visualization techniques

Text mining:

• Text mining, also known as text analysis, involves the extraction of valuable and significant information from unstructured text through an automated process. This method utilizes natural language processing techniques

This is a summary of the Chapter "Introduction to Social Network Analysis" (Maryam Amir Haeri and Stéphanie M. van den Berg)

These are content for the exam of the Research Methods component of Module 7 (Research Methods and Research Project) of the University of Twente, in the Netherlands. 

Introduction to Social Network Analysis

• Social network analysis (SNA) involves the study of social structures by incorporating network and graph theory --> it investigates network structures in terms of nodes (individual actors, organisations, or other types of objects within the network) and the ties or links that connect these nodes.
  • Quantitative and qualitative analysis of social network
    data

The SNA pipeline

• Problem identification: understanding problem and different aspects
  • What is the main objective of data analysis?
  • What should be predicted?
  • Why is that important?
  • How well does the prediction have to be?
  • Do we need real-time results?
• Selection of data sources: identify which social network(s) is (are) more appropriate for our projects --> crucial to consider privacy of users and policies of data collection for each social
  network platform
• Data cleaning: data collected from social networks may include duplicate, corrupted, irrelevant and missing data
• Data transformation: we need structured data (e.g. adjacency matrices) --> prepare data for analysis
• Mining (analysis): select approach for the analysis
• Analysis of results: criteria and methods of visualisation can be used to
  examine and interpret the results of data mining models

Social network data

• Social network consists of:
  • Actors (users)
    • User data: e.g. demographic info of the actors
  • Connections (links) --> the most important one is the structure of the network.
  • The contents transmitted over these connections --> message
    topics, hashtags, sentiment, etcetera

Networks represented as graphs

• Network to refer to the structure of social relationships
• Graph as the mathematical representation of a network --> is a set of “objects” in which some pairs of the objects are in some way “related” --> objects correspond to mathematical abstractions called vertices or nodes
  
  • Adjacent: Two vertices (nodes) that are connected by an edge --> e.g. if node Barry has a phone call with the node Anna, then we call nodes Barry and Anna adjacent
    
    • A graph is therefore a collection of vertices V and a collection of edges E
  • Path: sequence of edges which joins a sequence of vertices that are all distinct --> the length of a path is equal to the number of edges in the path
  • Degree of a vertex is the number of edges connected to this