![Sample our Education journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5932/TOC-Subject-Sample-2021-Education.jpg"})
Abstract
In many parts of the world, shadow education has become a major enterprise. Such is the case of the countries of the former Eastern Bloc, including the Czech Republic, which is in scope of this article. The study analyses the Internet supply of private tutoring lessons in academic subjects and assesses the micro- and macro-factors influencing the offered lesson price as set by private tutors. Based on the quantitative content analysis of 2058 individual tutor profiles advertising online, the author found a very unequal distribution of tutors within the country and a considerable proportion of mainstream schoolteachers acting as private lesson providers. The higher the formal education and age of the tutor, the higher the price they offer per lesson. Male tutors set higher prices than female tutors, and tutoring in foreign languages is generally more expensive than in any other subject. The results suggest that the features of private tutoring supply have implications for equity in education.
Notes
1. The terms ‘notice board’ and ‘mediated notice board’ are described in more detail by Tanner et al. (Citation2009).
2. Key words used for searching were: private tutor, private lessons, supplementary tutoring, help with academic subjects, remedial teaching to schoolchildren.
3. Namely, www.naucim.cz, www.rychlavyuka.cz, www.doucuji.eu, www.domelie.cz, www.doucka.cz, www.bezvadoucovani.cz, www.tutoriva.com, www.doucme.cz.
4. A lesson’s duration was stated as 45 or 60 minutes; for the purpose of further analysis, all prices were recalculated to reflect a 60-minute lesson. If the tutor was offering a price range, the average price was calculated from the lower and upper limit. For example, if the offered price was 150–200 CZK per 60 minute, calculations were made with 175 CZK per 60 minute.
5. Before undertaking the inquiry, the author coped with similar issues encountered by Tanner et al. (Citation2009, 9–10) concerning the legal and ethical aspects of collecting and assembling data from websites and constructing a research database. According to Article 88 of the Czech Copyright Act (No. 121/Citation2000), a database defined as ‘collection of independent works, data, or other items arranged in a systematic or methodical manner and individually accessible by electronic or other means, irrespective of the form of the expression thereof’ is protected by the copyright law, and according to paragraph 5, Article 88, ‘… repeated and systematic extraction or re-utilisation … of the content of the database … is not permitted.’ As the information stored on mediated notice boards for tutors fits the aforesaid definition of a database, using the notice boards as an information source in the way described in the law would have been illegal. Fortunately, the law anticipates an exception to this rule in Article 92: ‘The right of the maker of the database made available by him is not infringed by the lawful user who extracts or re-utilises a substantial part of the content of the database … b) for scientific or educational purposes, if he indicates the source, within the scope justified by the pursued non-gainful purpose’ (emphasis added). As the author’s intentions were scientific and non-profit, he could pursue the investigation.
6. Some preliminary findings were previously presented during a poster session at the conference of the Czech Education Research Association (Šťastný and Terreros, Citation2015) Although the poster was prepared by two authors, the author of this study draws only from the part he was solely responsible for.
7. The Variance Inflation Factor (VIF) of macro-level variables ranged from 5 to 31 in the models. In the literature, no universal threshold for the VIF indicator has yet been agreed; one can mostly find that values such as 3.3, 5 or 10 indicate the existence of multicollinearity between variables (Kock and Lynn Citation2012). The multicollinearity problem in regression models can be solved in multiple ways, one of the possible solutions is to lower the number of variables entering the models and to not include them in calculations (Morrow-Howell Citation1994).