UEA Dissertation Proposal: Q Statistic with Constant Weights Analysis

This dissertation proposal outlines a study focused on investigating the Q statistic with constant weights for assessing heterogeneity in meta-analyses. The proposal begins with background information on meta-analysis and the importance of addressing heterogeneity, followed by the motivation for the research and clearly defined aims and objectives. The study aims to evaluate the importance of Q statistics with constant weights, examine mean differences, and explore the role of Q statistics in meta-analysis. A critical review of the literature provides context, while the methodology section details the qualitative research approach, data collection methods, and data analysis techniques, including thematic analysis. Ethical, social, and legal issues, along with risk analysis and a project work plan, are also addressed. The proposal concludes with a comprehensive reference list and a brief abstract summarizing the study's focus and methods.

DISSERTATION
PROPOSAL

TABLE OF CONTENTS
BACKGROUND AND MOTIVATION.........................................................................................1
Background..................................................................................................................................1
Motivation....................................................................................................................................1
AIM AND OBJECTIVES...............................................................................................................2
Aim..............................................................................................................................................2
Objectives....................................................................................................................................2
Research questions.......................................................................................................................2
CRITICAL REVIEW OF LITERATURE.......................................................................................2
PROPOSED METHODOLOGY.....................................................................................................5
Research type...............................................................................................................................5
Research design...........................................................................................................................5
Research philosophy....................................................................................................................6
Research Approach......................................................................................................................7
Data Collection............................................................................................................................7
Data Analysis...............................................................................................................................8
ANALYSIS OF ETHICAL, SOCIAL AND LEGAL ISSUES.......................................................8
RISK ANALYSIS...........................................................................................................................9
PROJECT WORK PLAN AND GANTT CHART.......................................................................11
REFERENCES..............................................................................................................................14

ABSTRACT
This dissertation proposal focuses upon investigation of Q statistic with constant weights
for assessing heterogeneity in meta-analyses. In order to achieve this aim importance of q statics
in constant weight is identified, evaluation of mean difference for constant weight in accessing
heterogeneity, role of q static in meta- analysis is identified, and approximation of Q statistic
with constant weights for assessing heterogeneity in meta-analysis is done. Research methods
that this dissertation proposal will focus upon are: qualitative research type, secondary data will
be collected, interpretivism research philosophy, inductive research approach and for this
thematic data analysis method will be used.
BACKGROUND AND MOTIVATION
Background
Meta- analysis is a kind of statistical combination of results gathered from two or more
than two separate studies. This analysis helps in bringing improvement within precision in results
posed by different studies (Bakbergenuly, Hoaglin and Kulinskaya, 2021). But if it is not
considered carefully then it can result in mislead variation of studies (heterogeneity).
Heterogeneity in meta-analysis refers to different kinds of variation in studies study outcomes.
For meta- analysis variation across studies is extremely important to be considered there are
various methods that can be used for measuring heterogeneity in meta- analysis. In most of the
studies heterogeneity is measured using Cochran’s Q. But this method does not allow reliable
investigation of studies and its causes. In order to bring accuracy in approximation of
heterogeneity in meta- analysis there are various other methods that can be focused upon. One of
the methods is conventional Q statistic (Van Lissa, 2017). This Q statistics method is used in
meta-analysis to measure overall study heterogeneity or homogeneity. But in these various
problems arise in meta- analysis as this method uses inverse-variance (IV) weights. In order to
overcome this issue, Q statistics with constant weights can be focused upon as this method only
uses studies with effective sample size (Bakbergenuly, Hoaglin and Kulinskaya, 2021). This
method helps in bringing approximation with the help of Q statistics with constant weight.
Because of this reason in this dissertation investigation of investigate Q statistic with constant
weights for assessing heterogeneity in meta-analyses will be focused upon for understanding
ways in which this method can bring precision and approximation.
1

Motivation
The main motive of carrying out this research in meta- analysis is to understand extent till
which sizes vary in meta- analysis. It is extremely important to assess heterogeneity in meta-
analysis so that accuracy within results can be brought. High heterogeneity in meta-analysis is
also a symbol that there are two or more than two sub-groups in present studies which is a bit
difficult to be true (Friede and et. al., 2017). This reason motivated me to carry out research in
this meta- analysis field for investigating Q statistic with constant weights for assessing
heterogeneity in meta-analyses.
AIM AND OBJECTIVES
Aim
Main of this dissertation proposal is ‘To investigate Q statistic with constant weights for
assessing heterogeneity in meta-analyses.’
Objectives
In order to achieve main aim of this project, following research objectives will be focused
upon and will be achieved by the research. Research objectives of this dissertation proposal that
will be achieved are as follows:
 To evaluate importance of q statics in constant weight.
 To examine mean difference for constant weight in accessing heterogeneity
 To examine role of q static in meta- analysis.
 To investigate approximation of Q statistic with constant weights for assessing heterogeneity
in meta-analysis
Research questions
In this dissertation following research questions will also be answered:
 What is the role of Q statics in constant weight?
 How important is mean difference for constant weight in accessing heterogeneity?
 What is the role of Q static in meta- analysis?
CRITICAL REVIEW OF LITERATURE
Importance of Q statistics in constant weight
According to Tong and Guo, (2019), the q statistics is basically consider the conceptual
termed in which used for multiple significance testing across number of means. For Example- x-
>y, this statistics is related to mathematical termed. In order to identify the result, test the
2

modified version which provide better sample properties. When evaluate the expression, it is
always important aspect in order to done within evaluation mode, either mid or sample mode. As
a result, it can easily computed by expression which may different in two models but at certain
place, it would be increase uncertainty. The Q statistics server as two main phases in context of
random effects meta- analysis testing for presence of heterogeneity, estimating or calculating the
different between study variances.
Bowden, Hemani and Davey (2018) The importance of q statistics is basically
represented routinely used for purpose of testing heterogeneity in meta-analysis. In certain level,
it is expected the value in which help for estimating the comparison between different study
variance. Different kind of applications are generally not applicable as implication of use of
estimated variances in context of inverse-variance weights. The most importantly, it may be
calculated the weights which make approximating the distribution Q, rather it is very
complicated. Author said that identified the significance of Q statistic at the time of random
effect model. For study, it represent the value (i=1,2,......, k) estimate the effect of ^_i _ G(_i;
v2i ), where it can be identified or measured the specific distribution G, it has mean of _i and
variance v2i , and _i _ N(_; _ 2). Thus, it has been identified the unbiased estimates of true
condition. Q statistic is basically gained some weighted sum of its squared deviations of its
estimated effect (Bakbergenuly, Hoaglin and Kulinskaya, 2021). On the basis of its explanation,
it has been identified the importance of q statistics in the calculation of equations.
Role of q statics in meta-analysis
As per Ren, Oakley and Stevens (2018), identified the role of q statistics in meta-
analysis, which means that assume an appropriate equations and exists in the form of quadratic.
But it should be used the vector has used the multivariate normal distribution, Q is generally
obtained through algorithm in order to determine the eigenvalues of A sigma and applicable
some other kind of inputs. In case if variances exist in sigma are true variances. According to the
farebrother’s algorithum, it has been evaluated the exact distribution of Q values. In practices, it
is an essential aspect in which plug in estimated variances. Afterwards, it is becoming easy to
find the result as per approximation and which quite accurate for MD.
Author expressed their feeling towards q statistics, which is mainly used to measure,
predict or calculate homogeneity in meat analysis. Basically, it can be divided into different ways
to think about set of various studies. Afterwards, it become easier for analyzing the Meta data or
3

information. All kind of studies are capturing the exact same kind of effects, so it is very tough
for differentiating the between q statistics in meta- analysis. Generally, it has been drawn the
sample from distribution of effects over and estimate the particular result or outcome. On the
basis of author, it is not true that capture or find same kind of effect. At some point, it may
identify the difference in values because of random variations. It could be occurred due to
difference in the effect that stud is capturing (Bakbergenuly, Hoaglin and Kulinskaya, 2021). The
Q statistics is basically used to try the partition in variation because it can be measured on the
basis of previous studies. The variability is all about the effect in which estimate or calculate
potential differences between multiple studies. That’s why, in meta-analysis it has been used the
q statistics in which identify the variations within different values.
Understand the concept of meta analysis in Constant weight in accessing heterogeneity.
As per Kossmeier, Tran and Vorace (2020), the concept of meta- analyses means that
provide accurate reliable evidence relevant to different aspects in context of estimation.
Especially, it is clear way to identify the effect of similar magnitudes. Meta-analyses is basically
represent the statistics test of heterogeneity. This kind of test seeks to identify whether there is
genuine differences underlying in result or outcome. A test of heterogeneity is basically
identified the null hypothesis that mainly indicate the same kind of effects. Usually, different
kind of test statistics are computed by considerations of squared deviations and also estimate
from entire metal-analyses.
The weighting to constant weight, its purpose is to identify the sample values and ensure
that focus on the specific mass of sample. But it is important that does not include a variable
amount otherwise, it may show a variations (Bakbergenuly, Hoaglin and Kulinskaya, 2021).
During meta-analyses, constant weight can be assumed in q statistics in which remove or
eliminate the variations and then estimated the similar values. Whenever, accessing the
heterogeneity which means that examine the variations but it can be changed into constant
weight but needs to be include additional values approximately.
As per Oakley and Stevens (2018), views, Meta –analysis is one of the best statistical
combination of results from different separate studies. The potential benefit or advantage is to
determine the importance of assessing heterogeneity. The ability is to find the answer of
particular questions (Bakbergenuly, Hoaglin and Kulinskaya, 2021). At certain point, it will be
increasing opportunity to settle controversies arising from conflicts. However, it may have been
4

increasing the potential mislead seriously, this kind of situation is basically occurred in design,
study biases, variations across studies. Afterwards, it is becoming easily examine the reporting
biases which are not consider at the time of estimation.
Furthermore, variation across heterogeneity must be considered, although it reviews the
enough studies and allow them to investigate on the significant causes. Because the random
effects identified in the meta- analyses allows for heterogeneity by considerations that underlying
effects follow as normal distribution. this is the best way to interpret data carefully, prediction
intervals from random effect and also useful device for representing the extent of between study
variations.
PROPOSED METHODOLOGY
Research type
Research type are kind of methods that helps in describing type of data that will be
collected and analysed in the research. It can also be defined as a kind of systematic approach of
collection data, analysing, interpretation in order to achieve main research aim and answer main
research questions in an appropriate manner (Snyder, 2019). Research type can be classified into
three main types of research that are: qualitative, quantitative and mixed research type. It is
extremely important to select appropriate research type as per type of research and type of aim
and objectives that are required to be achieved. First type of research is qualitative research type.
In this type of research, data collected is non-numerical or in form of word, emotions, feelings,
sounds, emotions or are unquantifiable elements. All the information or data collected in this
type of research is qualitative that cannot be analysed using mathematical methods or formula.
Another type of research quantitative research type. It is a type of research in which main
research problem is resolved sing numbers or numerical data which is analysed using statistical
analysis method. Last type of research is mixed research type. It is a kind of research in which
both qualitative and quantitative data are collected and analysed. In this type of research both
numerical or statistical and qualitative methods are used for analysis of numerical and non-
numerical data collected.
For this dissertation qualitative research type will be focused upon because in this
dissertation previous studies of (Bakbergenuly, and et. al., 2021), and (Kulinskaya and et. al.,
2021) will be analysed for achievement of main aim of this research.
5

Research design
Research deign can be explained as a kind of framework of different research methods
and techniques that are chosen by researcher or achievement of main objectives of research.
Choice of research design completely depends upon type of research which is being carried out,
type of data that will be collected by researcher and type of analysis method that will be chosen
by the researcher (Ørngreen and Levinsen, 2017). There are three main type of research designs,
that are: exploratory, descriptive and casual research design. First type of research design is
exploratory research design that focuses upon exploring specific aspect of research area so that
main research questions of research can be answered in an appropriate manner. this type of
research design is chosen when open ended data is collected that do not have a fixed set of
procedure. Next research deign is descriptive research design. It is a kind of research design in
which description of specific causes, elements, and phenomenon of a research is done. Selection
of this research design is used when researcher do not know much about the research area or
topic. Last type of research design is casual research design. It is a kind of design that is used
when there is a need to study cause and effect relationship within research. When there is a need
to analyse a situation or find out patter between two or more than two variables then this design
is used.
For this research exploratory research design will be used as it is a qualitative research in
which exploration of metal analysis research area will be done for answering main research
questions.
Research philosophy
Research philosophy can be explained as a kind of belief about the way in which data
bout a phenomenon should be gathered, used and analysed. The main reason behind selecting an
appropriate research philosophy is to choose appropriate data collection method. There are three
main types of research philosophies that can be used by researcher, that are: interpretivism,
positivism, and pragmatism (Mohajan, 2018). First philosophy is interpretivism research
philosophy which is used to interpret elements of a research study in order to integrate human
interest into the study. This philosophy is used to explain that access to reality is only through
social constructs. Whereas positivism research philosophy says that science is the only way to
learn the truth. It is used when there is only requirement of gaining factual knowledge through
observations. Third type of philosophy is pragmatism philosophy, which says that a concept can
6

only be accepted if it support actions. This philosophy says that there are various methods of
interpreting the world and undertaking a research because of which no single point of view can
provide entire picture of reality.
As it is a qualitative research in which secondary data will be collected so in this
interpretivism research philosophy will be used by researcher and using this data human interest
will be integrated into the research.
Research Approach
Research approach can be explained as a systematic piece of work undertaken to increase
knowledge stock. It is a kind of procedure or plan for research that provide broad assumption of
detailed methods of data collection, analysis and interpretation. There are two main types of
research approaches that can be used by researcher, that are: indictive and deductive research
approaches (Zangirolami-Raimundo, Echeimberg and Leone, 2018). Inductive research approach
is a kind approach theories and observations are proposed at the end of research process though
observation results. In this a pattern from observation is drawn and developed so that formulated
research questions can be answered (Kumar, 2018). Another research approach is deductive
approach. In this approach existing theories are used for development of hypothesis and then on
the basis of those hypothesis research strategy is designed for testing hypothesis. It is simply
known as reasoning from particular to general. Using this approach casual relationship is applied
to a particular theory that might be true in many cases. It is important for a researcher to choose
an appropriate research approach because it can help researcher to choose and involve
philosophical assumptions as well as assume appropriate method or procedure that can be
followed by researcher for achievement of main aim and objectives of the research.
For this dissertation researcher will be using inductive research approach as it is a
qualitative study adoption of inductive approach will directly help researcher in drawing
important pattern from observation so that main research questions can be answered in an
appropriate manner.
Data Collection
Data collection can be explained as a procedure of collecting, analysing and measuring
accurate insight for research using different kinds of validation techniques. Data collection is one
of the most important part of research or dissertation that helps the researcher in finding
appropriate and correct answers to main research questions (Ngozwana, 2018). Adoption of an
7

appropriate data collection method completely depends upon type of research which is being
conducted and aim and objectives that are required to be achieved. There are two main types of
data that can be collected by researcher that are: primary and secondary data. Method of
collecting both primary and secondary data are completely different. First type of data primary
data, it is a kind of data which is collected by researcher for the first time at the time of research
(Abutabenjeh and Jaradat, 2018). This type of data is collected either though interview,
questionnaire, survey, observations and many more. Whereas, another type of data is secondary
data. It is a kind of data that has already been published in books, journals, articles, newspapers,
etc. This data is collected from online portals, government authorised websites, newspapers,
articles, research studies, journals and many more. It is important to choose an appropriate type
of data and method though which it will be collected because it directly impacts achievement of
main aim and objectives of research.
For this dissertation, researcher will be collecting secondary data in which this data will
be collected from already published articles, research studies and online portals.
Data Analysis
Data analysis can be defined as a kind of process of inspecting, transforming, cleaning and
modelling data in order to transform it into some useful information that can be used to wither
answer main research questions or achieve main objectives of the research. There are two main
types of data analysis methods, that are thematic and statistical methods. Thematic analysis is a
qualitative analysis method which is used for qualitative studies for analysis of non-numerical
data such as text, interviews etc. Another type of data analysis method is statistical data analysis
method (Basias and Pollalis, 2018). This method is used for uncovering trends and patterns
though analysis of numerical data for extraction of important information can be calculated with
this analysis method. Statistical analysis method is mostly used with quantitative studies that
consist of quantitative data or numerical data. Choice of data can analysis method completely
depend upon type of research which is being carried out and type of data which is being
collected because type of analysis method chosen for interpretation and analysis directly impact
results obtained and research questions that are required to be answered.
In this dissertation, researcher will be adopting thematic data analysis method as in this
research secondary data is being collected by the researcher and thematic data analysis method is
one of the most appropriate analysis method that can be opted by the researcher for analysis of
8

secondary data that will be collected by the researcher and will further help in investigating and
analysing approximation of Q statistic with constant weights for assessing heterogeneity in meta-
analysis. For analysis of data, algorithms will be used for simulation such as LOG ODD RATIO,
LOR Q-fixed for LOR, Q statistics estimation for heterogeneity for meta analysis, mean
difference, standard mean difference in q statistics. Each of these algorithms have their own use
that can help in creation of simulation program. Log Odd ration is very useful for solving certain
problems, basically ones related to finding probabilities. Whereas Q statistics estimation for
heterogeneity is used for meta analysis. Q statistics can be used for calculating both homogeneity
and heterogeneity by calculating difference in studies due to variation can be calculated with the
help of this Q statistics estimation. Lastly mean difference will also be used in this for creation of
simulation program as it will help in measuring the absolute difference between the mean value
in two groups in a clinical trial.
The implementation of Simulation program
The simulation implementation in context of Q Statistics in Meta data Analysis. in order
to require an introduce inconsistency in simulated network by multiplying consistent odds ratio
like OR, ration of odds Ratios ROR.
This is because of high level of computation in context of bayesian approaches, it is always kept
all different paramters. On the basis of analysis, it would be doing some comparison in the
network, where wach contained specific data.
Following different steps of simulation data implementation and program running-
Simply do comparison
Draw i= 1,,,k
Specific log odds ration
Implement formulas :-
9

Afterwards, it was randomly generated binomial distribution with specific paramter n. for all
pariwide comparsion intervention c with out consideration of inconsistency.
ANALYSIS OF ETHICAL, SOCIAL AND LEGAL ISSUES
This project does not involve any legal issues or other social issues as in this dissertation
no primary research is being carried out but, in this research, some ethical issues can be arise in
this dissertation project. Some of the main ethical issues that can arise in this dissertation project
and can be faced by the researcher are as follows: non adherence of ‘Data protection Act, 2018’
can result in arising ethical issues that can be faced by the researcher in this dissertation. Another
ethical issue that can arise in this dissertation is non-checking authenticity of secondary data
collected (Dodds and Hess, 2020). Usage of secondary data without checking authenticity of the
data, whether it has been published by an authorised publisher or not. It is an ethical issue that
might arise in this dissertation project. So in order to avoid this issue it is important to check
authenticity of secondary information and to check whether it is published by authorised
publisher or not In order to avoid this ethical issue secondary data can be collected from
authorised libraries or databases such as Google scholar, Emerald database. Another issue that
can arise due in this dissertation project is ethical obligation of usage of secondary data i.e.,
using other researcher’s data without their consult or without citing main source from where that
data had been collected in an appropriate manner.
RISK ANALYSIS
Some of the major risk or potential threat that might arise in this dissertation project that
might put this project at risk of not being completed on time have been explained below with
contingency plan that can help in eliminating those problems or irks are as follows:
Risk Likelihood Impact Contingency plan
10